rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
#!/usr/bin/env python3
|
|
|
|
|
2021-06-01 20:35:42 +00:00
|
|
|
# Copyright Red Hat
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
#
|
|
|
|
# createhdds is free software; you can redistribute it and/or modify
|
|
|
|
# it under the terms of the GNU General Public License as published by
|
|
|
|
# the Free Software Foundation, either version 3 of the License, or
|
|
|
|
# (at your option) any later version.
|
|
|
|
#
|
|
|
|
# This program is distributed in the hope that it will be useful,
|
|
|
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
# GNU General Public License for more details.
|
|
|
|
#
|
|
|
|
# You should have received a copy of the GNU General Public License
|
|
|
|
# along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
#
|
|
|
|
# Author: Adam Williamson <awilliam@redhat.com>
|
|
|
|
|
|
|
|
"""Tool for creating hard disk images for Fedora openQA."""
|
|
|
|
|
|
|
|
import argparse
|
|
|
|
import logging
|
|
|
|
import json
|
|
|
|
import os
|
2021-06-18 21:52:14 +00:00
|
|
|
import os.path
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
import subprocess
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
import sys
|
|
|
|
import tempfile
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
import time
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
|
|
|
import fedfind.helpers
|
|
|
|
import guestfs
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
import libvirt
|
2017-02-20 16:08:19 +00:00
|
|
|
import platform
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
|
|
|
from six.moves.urllib.request import urlopen
|
|
|
|
|
|
|
|
# this is a bit icky, but it means you can run the script from
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
# anywhere - we use this to locate hdds.json and the virtinstall
|
|
|
|
# image kickstarts, so they must always be in the same place
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# as the script itself. images are checked/created in the working
|
|
|
|
# directory.
|
|
|
|
SCRIPTDIR = os.path.abspath(os.path.dirname(sys.argv[0]))
|
2017-02-20 16:08:19 +00:00
|
|
|
CPUARCH = platform.processor()
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
logger = logging.getLogger('createhdds')
|
|
|
|
|
|
|
|
def handle_size(size):
|
|
|
|
"""Simple function to handle sizes like '10G' or '100MB', returns
|
|
|
|
the size in bytes as an int. Used by both image classes.
|
|
|
|
"""
|
|
|
|
size = str(size)
|
|
|
|
if size.endswith('G') or size.endswith('GB') or size.endswith('GiB'):
|
|
|
|
return int(size.split('G')[0]) * 1024 * 1024 * 1024
|
|
|
|
elif size.endswith('M') or size.endswith('MB') or size.endswith('MiB'):
|
|
|
|
return int(size.split('M')[0]) * 1024 * 1024
|
|
|
|
else:
|
|
|
|
return int(size)
|
|
|
|
|
2017-08-18 00:42:43 +00:00
|
|
|
def supported_arches():
|
|
|
|
"""Provides a list of the arches for which virt-install images can
|
|
|
|
be built on this host.
|
|
|
|
"""
|
2017-10-16 16:51:56 +00:00
|
|
|
powerpc_arches = ['ppc64', 'ppc64le', 'noarch']
|
|
|
|
intel_arches = ['i686', 'x86_64', 'noarch']
|
2018-03-06 17:06:58 +00:00
|
|
|
aarch64_arches = ['aarch64', 'armv7l']
|
2017-08-18 00:42:43 +00:00
|
|
|
if CPUARCH in powerpc_arches:
|
|
|
|
supported_arches = powerpc_arches
|
|
|
|
elif CPUARCH in intel_arches:
|
|
|
|
supported_arches = intel_arches
|
2018-03-06 17:06:58 +00:00
|
|
|
elif CPUARCH == 'aarch64':
|
|
|
|
supported_arches = aarch64_arches
|
2017-08-18 00:42:43 +00:00
|
|
|
else:
|
|
|
|
supported_arches = []
|
|
|
|
logger.info("Need to add a list of supported arches for %s CPU", CPUARCH)
|
|
|
|
return supported_arches
|
|
|
|
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
class GuestfsImage(object):
|
|
|
|
"""Class representing an image created by guestfs. 'size' is the
|
|
|
|
desired image size, valid formats are a digit string (size in
|
|
|
|
bytes, digit string plus 'M', 'MB' or 'MiB' (size in power of two
|
|
|
|
megabytes), or digit string plus 'G', 'GB' or 'GiB' (size in power
|
|
|
|
of two gigabytes). 'imgver' is the image 'version' - in practice
|
|
|
|
it's simply a string that gets included in the image file name
|
|
|
|
if specified. 'filesystem' is the default filesystem for the image
|
|
|
|
- it will be used for parts that don't explicitly specify a
|
|
|
|
filesystem. 'label' is the disk label format to be used. parts,
|
|
|
|
writes, and uploads are lists of dicts that specify the partitions
|
|
|
|
that should be created and files that should be written or copied
|
|
|
|
('uploaded') to them. These are read in from hdds.json by
|
|
|
|
get_guestfs_images and passed here. 'name_extras' is an iterable
|
|
|
|
of strings which should be appended to the image file name, with _
|
|
|
|
separators (this provides a mechanism for get_guestfs_images to
|
|
|
|
include the label and/or filesystem in the image name when
|
|
|
|
creating images from a group with variants).
|
|
|
|
"""
|
|
|
|
def __init__(self, name, size, imgver='', filesystem='ext4', label='mbr', parts=None,
|
|
|
|
writes=None, uploads=None, name_extras=None):
|
|
|
|
self.name = name
|
2017-10-16 16:51:56 +00:00
|
|
|
self.arch = 'noarch'
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
self.size = handle_size(size)
|
|
|
|
# guestfs images are never outdated
|
|
|
|
self.outdated = False
|
|
|
|
self.filesystem = filesystem
|
|
|
|
self.label = label
|
|
|
|
self.parts = []
|
|
|
|
self.writes = []
|
|
|
|
self.uploads = []
|
|
|
|
if parts:
|
|
|
|
self.parts = parts
|
|
|
|
if writes:
|
|
|
|
self.writes = writes
|
|
|
|
if uploads:
|
|
|
|
self.uploads = uploads
|
|
|
|
|
|
|
|
self.filename = "disk_{0}".format(name)
|
|
|
|
if imgver:
|
|
|
|
self.filename = "{0}_{1}".format(self.filename, imgver)
|
|
|
|
if name_extras:
|
|
|
|
for item in name_extras:
|
|
|
|
self.filename = "{0}_{1}".format(self.filename, item)
|
|
|
|
self.filename = "{0}.img".format(self.filename)
|
|
|
|
|
2017-02-15 01:41:31 +00:00
|
|
|
def create(self, _):
|
|
|
|
"""Create the image. The unused arg is the 'textinst' arg that
|
|
|
|
only VirtInstallImages care about (but which has to be passed
|
|
|
|
here too).
|
|
|
|
"""
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
gfs = guestfs.GuestFS(python_return_dict=True)
|
|
|
|
try:
|
|
|
|
# Create the disk image with a temporary name
|
|
|
|
tmpfile = "{0}.tmp".format(self.filename)
|
|
|
|
gfs.disk_create(tmpfile, "raw", int(self.size))
|
|
|
|
# 'launch' guestfs with the disk attached
|
|
|
|
gfs.add_drive_opts(tmpfile, format="raw", readonly=0)
|
|
|
|
gfs.launch()
|
|
|
|
# identify the disk and create a disk label
|
|
|
|
disk = gfs.list_devices()[0]
|
|
|
|
gfs.part_init(disk, self.label)
|
|
|
|
# create and format the partitions
|
|
|
|
for part in self.parts:
|
|
|
|
# each partition can specify a filesystem, if it doesn't,
|
|
|
|
# we use the image default
|
|
|
|
if 'filesystem' not in part:
|
|
|
|
part['filesystem'] = self.filesystem
|
|
|
|
# create the partition: the dict must specify type ('p'
|
|
|
|
# for primary, 'l' for logical, 'e' for extended), and
|
|
|
|
# start and end sector numbers - more details in
|
|
|
|
# guestfs docs
|
|
|
|
gfs.part_add(disk, part['type'], int(part['start']), int(part['end']))
|
2019-10-18 10:03:37 +00:00
|
|
|
# identify the partition
|
|
|
|
partname = gfs.list_partitions()[-1]
|
|
|
|
partnum = gfs.part_list(disk)[-1]["part_num"]
|
|
|
|
# sometimes, we want to set the gpt type of the partition
|
|
|
|
gpt_type = part.get("gpt_type", None)
|
|
|
|
if gpt_type:
|
|
|
|
gfs.part_set_gpt_type(disk, partnum, gpt_type)
|
|
|
|
# format the partition
|
|
|
|
gfs.mkfs(part['filesystem'], partname, label=part.get('label'))
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# do file 'writes' (create a file with a given string as
|
|
|
|
# its content)
|
|
|
|
for write in self.writes:
|
|
|
|
# the write dict must specify the partition to be
|
|
|
|
# written to, numbered from 1 as humans usually do;
|
|
|
|
# find that part and mount it
|
|
|
|
partn = gfs.list_partitions()[int(write['part'])-1]
|
|
|
|
gfs.mount(partn, "/")
|
|
|
|
# do the write: the dict must specify the target path
|
|
|
|
# and the string to be written ('content')
|
|
|
|
gfs.write(write['path'], write['content'])
|
2016-05-04 18:55:22 +00:00
|
|
|
gfs.sync()
|
|
|
|
gfs.umount_opts("/")
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# do file 'uploads'. in guestfs-speak that means transfer
|
|
|
|
# a file from the host to the image, we use it to mean
|
|
|
|
# download a file from an http server and transfer that
|
|
|
|
# to the image
|
|
|
|
for upload in self.uploads:
|
2016-03-21 15:53:13 +00:00
|
|
|
# as with write, the dict must specify a target
|
|
|
|
# partition and location ('target')
|
|
|
|
partn = gfs.list_partitions()[int(upload['part'])-1]
|
|
|
|
gfs.mount(partn, "/")
|
keep upload files locally, activate network in kickstarts
Summary:
I've come to dislike the approach where we source files that
get included in the disk images remotely; it's unnecessarily
complicated and needlessly exposes us to unexpected changes in
the remote files. So simply keep the files in this git repo
instead, none of them is very big. This also simplifies the
code.
Also, change all the kickstarts to specify --device=link and
--activate on their network lines. These arguments *should*
be included, according to the documentation (the first to
specify which device the config is for, the second to specify
that it should be activated in the installer environment), and
I think the lack of the second is actually now a problem for
the FreeIPA kickstart enrolment test (not sure exactly what
changed to make this a problem where it wasn't before, but
*something* has, and this fixes it).
Test Plan:
check that all tests still run properly. The
FreeIPA enrolment kickstart test should now get somewhat
further. Before this change, for F25 and Rawhide, it fails to
enrol at all during install. Now it should enrol properly -
that's what the kickstart changes fix, it was failing because
it wasn't using the right network config - but it still fails
when trying to log in as test1, due to RHBZ #1366403 .
Reviewers: jskladan, garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D968
2016-08-16 00:02:05 +00:00
|
|
|
gfs.upload('/'.join((SCRIPTDIR, 'uploads', upload['source'])), upload['target'])
|
2016-05-04 18:55:22 +00:00
|
|
|
gfs.sync()
|
|
|
|
gfs.umount_opts("/")
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# we're all done! rename to the correct name
|
|
|
|
os.rename(tmpfile, self.filename)
|
|
|
|
except:
|
|
|
|
# if anything went wrong, we want to wipe the temp file
|
|
|
|
# then raise
|
|
|
|
os.remove(tmpfile)
|
|
|
|
raise
|
|
|
|
finally:
|
|
|
|
# whether things go right or wrong, we want to close the
|
2016-03-21 15:53:13 +00:00
|
|
|
# gfs instance, and rwmj recommends 'shutdown()' too
|
|
|
|
gfs.shutdown()
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
gfs.close()
|
|
|
|
|
|
|
|
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
class VirtInstallImage(object):
|
|
|
|
"""Class representing an image created by virt-install. 'release'
|
|
|
|
is the release the image will be built for. 'variant' is the
|
|
|
|
variant whose install tree should be used. 'arch' is the arch.
|
|
|
|
'size' is the desired image size, in gigabytes. 'imgver' is
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
the image 'version' - in practice it's simply a string that gets
|
|
|
|
included in the image file name if specified. 'maxage' is the
|
|
|
|
maximum age of the image file (in days) - if the image is older
|
|
|
|
than this, 'check' will report it as 'outdated' and 'all' will
|
2019-11-12 14:00:47 +00:00
|
|
|
rebuild it. 'bootopts' are used to pass boot options to the
|
2022-04-26 16:59:23 +00:00
|
|
|
virtual image to provide better control of the VM. 'branched'
|
|
|
|
and 'rawhide' are flags for whether this is a branched or Rawhide
|
|
|
|
release (which affects the paths).
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
"""
|
2022-05-25 21:54:25 +00:00
|
|
|
def __init__(self, name, release, arch, size, variant=None, imgver='', maxage=7,
|
2022-04-26 16:59:23 +00:00
|
|
|
bootopts=None, branched=False, rawhide=False):
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
self.name = name
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
self.size = size
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
self.filename = "disk_f{0}_{1}".format(str(release), name)
|
|
|
|
if imgver:
|
|
|
|
self.filename = "{0}_{1}".format(self.filename, imgver)
|
2021-09-10 17:24:40 +00:00
|
|
|
self.filename = "{0}_{1}.qcow2".format(self.filename, arch)
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
self.release = release
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
self.variant = variant
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
self.arch = arch
|
|
|
|
self.maxage = maxage
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
if variant:
|
|
|
|
self.variant = variant
|
|
|
|
else:
|
2017-02-15 01:41:31 +00:00
|
|
|
if str(release).isdigit() and int(release) < 24:
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
self.variant = "Server"
|
|
|
|
else:
|
|
|
|
self.variant = "Everything"
|
2019-11-12 14:00:47 +00:00
|
|
|
self.bootopts = bootopts
|
2022-04-26 16:59:23 +00:00
|
|
|
self.rawhide = rawhide
|
|
|
|
self.branched = branched
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
2021-06-18 21:52:14 +00:00
|
|
|
@property
|
|
|
|
def kickstart_file(self):
|
|
|
|
"""Find the most specific kickstart file to use for a given
|
|
|
|
name, release and arch. Order of preference:
|
|
|
|
* name-release-arch.ks
|
|
|
|
* name-release.ks
|
|
|
|
* name-arch.ks
|
|
|
|
* name.ks
|
|
|
|
"""
|
|
|
|
cands = [
|
|
|
|
f"{self.name}-{self.release}-{self.arch}.ks",
|
|
|
|
f"{self.name}-{self.release}.ks",
|
|
|
|
f"{self.name}-{self.arch}.ks",
|
|
|
|
f"{self.name}.ks"
|
|
|
|
]
|
|
|
|
for cand in cands:
|
|
|
|
if os.path.isfile("/".join((SCRIPTDIR, cand))):
|
|
|
|
logger.debug("Using kickstart %s", cand)
|
|
|
|
return cand
|
|
|
|
|
2017-02-15 01:41:31 +00:00
|
|
|
def create(self, textinst, retries=3):
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
"""Create the image."""
|
2017-08-18 00:42:43 +00:00
|
|
|
if self.arch not in supported_arches():
|
2018-03-06 16:54:35 +00:00
|
|
|
logger.info("Won't create %s image on %s host. This is normal, don't worry. If you "
|
|
|
|
"intend to have %s workers you will need to run createhdds again on one "
|
2018-03-15 17:35:13 +00:00
|
|
|
"of them to create their base images", self.arch, CPUARCH, self.arch)
|
2017-08-18 00:42:43 +00:00
|
|
|
return
|
|
|
|
|
2017-02-15 01:41:31 +00:00
|
|
|
# figure out the best os-variant. NOTE: libosinfo >= 0.3.1
|
|
|
|
# properly returns 1 on failure, but using workaround for old
|
|
|
|
# bug where it didn't in case EPEL doesn't have 0.3.1
|
|
|
|
shortid = "fedora{0}".format(self.release)
|
|
|
|
args = ["osinfo-query", "os", "short-id={0}".format(shortid)]
|
|
|
|
process = subprocess.Popen(args, stdout=subprocess.PIPE, stderr=subprocess.DEVNULL)
|
|
|
|
out = process.communicate()[0].decode()
|
|
|
|
if shortid not in out:
|
|
|
|
# this will just use the most recent fedora release number
|
|
|
|
# virt-install / osinfo knows
|
|
|
|
shortid = 'fedora-unknown'
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
|
|
|
|
# destroy and delete the domain we use for all virt-installs
|
|
|
|
conn = libvirt.open()
|
|
|
|
try:
|
|
|
|
dom = conn.lookupByName('createhdds')
|
|
|
|
try:
|
|
|
|
dom.destroy()
|
|
|
|
except libvirt.libvirtError:
|
|
|
|
# domain may not be running, so this is fine
|
|
|
|
pass
|
2018-03-14 20:42:01 +00:00
|
|
|
dom.undefineFlags(libvirt.VIR_DOMAIN_UNDEFINE_NVRAM)
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
except libvirt.libvirtError:
|
|
|
|
# domain may not exist, so this is fine
|
|
|
|
pass
|
|
|
|
conn.close()
|
|
|
|
|
2016-05-23 13:55:17 +00:00
|
|
|
tmpfile = "{0}.tmp".format(self.filename)
|
2017-07-19 20:08:31 +00:00
|
|
|
arch = self.arch
|
2017-09-04 23:06:44 +00:00
|
|
|
fedoradir = 'fedora/linux'
|
2021-09-10 16:19:13 +00:00
|
|
|
memsize = '3072'
|
2017-09-04 19:18:40 +00:00
|
|
|
if arch == 'i686':
|
|
|
|
arch = 'i386'
|
|
|
|
if arch in ['ppc64','ppc64le']:
|
|
|
|
fedoradir = 'fedora-secondary'
|
|
|
|
memsize = '4096'
|
2019-11-06 19:58:18 +00:00
|
|
|
if arch == 'i386':
|
|
|
|
# i686 is in fedora-secondary (until it died)
|
2017-09-04 23:06:44 +00:00
|
|
|
fedoradir = 'fedora-secondary'
|
2017-09-04 19:18:40 +00:00
|
|
|
|
keep variant setting removed by previous commit
to avoid error like:
===
$~/createhdds/createhdds.py support
...
Traceback (most recent call last):
File "/root/createhdds/createhdds.py", line 821, in <module>
main()
File "/root/createhdds/createhdds.py", line 811, in main
args.func(args, hdds)
File "/root/createhdds/createhdds.py", line 701, in cli_image
img.create(args.textinst)
File "/root/createhdds/createhdds.py", line 291, in create
loctmp.format(fedoradir, str(self.release), variant, arch), "--name", "createhdds",
NameError: name 'variant' is not defined
===
Signed-off-by: Michel Normand <normand@linux.vnet.ibm.com>
2019-02-06 08:31:59 +00:00
|
|
|
variant = self.variant
|
2019-08-27 15:15:56 +00:00
|
|
|
# From F31 onwards, Workstation tree is not installable and we
|
|
|
|
# build Workstation images out of Everything
|
2022-04-22 20:39:41 +00:00
|
|
|
if variant == 'Workstation':
|
2019-08-27 15:15:56 +00:00
|
|
|
variant = 'Everything'
|
2019-11-12 14:00:47 +00:00
|
|
|
|
2016-05-23 13:55:17 +00:00
|
|
|
try:
|
2017-02-15 01:41:31 +00:00
|
|
|
# this is almost complex enough to need fedfind but not
|
|
|
|
# quite, I think. also fedfind can't find the 'transient'
|
|
|
|
# rawhide and branched locations at present
|
2022-04-26 16:59:23 +00:00
|
|
|
if self.rawhide:
|
2017-09-04 19:18:40 +00:00
|
|
|
loctmp = "https://dl.fedoraproject.org/pub/{0}/development/rawhide/{2}/{3}/os"
|
2022-04-26 16:59:23 +00:00
|
|
|
elif self.branched:
|
2017-09-16 02:56:35 +00:00
|
|
|
loctmp = "https://dl.fedoraproject.org/pub/{0}/development/{1}/{2}/{3}/os/"
|
2017-02-15 01:41:31 +00:00
|
|
|
else:
|
2019-11-06 19:58:18 +00:00
|
|
|
loctmp = "https://download.fedoraproject.org/pub/{0}/releases/{1}/{2}/{3}/os/"
|
2021-06-18 21:52:14 +00:00
|
|
|
ksfile = self.kickstart_file
|
2021-05-13 23:22:24 +00:00
|
|
|
xargs = "inst.ks=file:/{0}".format(ksfile)
|
2022-04-22 23:42:24 +00:00
|
|
|
if str(self.release) == '35':
|
|
|
|
# we need this for installs to work with updated systemd:
|
|
|
|
# https://bugzilla.redhat.com/show_bug.cgi?id=2019579#c24
|
|
|
|
xargs += " inst.updates=https://rvykydal.fedorapeople.org/update-images/updates.f35-2019579-resolvconf.img"
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
args = ["virt-install", "--disk", "size={0},path={1}".format(self.size, tmpfile),
|
2017-04-08 03:14:33 +00:00
|
|
|
"--os-variant", shortid, "-x", xargs, "--initrd-inject",
|
2021-05-13 23:22:24 +00:00
|
|
|
"{0}/{1}".format(SCRIPTDIR, ksfile), "--location",
|
2017-09-04 19:18:40 +00:00
|
|
|
loctmp.format(fedoradir, str(self.release), variant, arch), "--name", "createhdds",
|
2019-11-12 14:00:47 +00:00
|
|
|
"--memory", memsize, "--noreboot", "--wait", "-1"]
|
2020-05-09 16:13:19 +00:00
|
|
|
if logger.getEffectiveLevel() == logging.DEBUG:
|
|
|
|
# let's get virt-install debug logs too
|
|
|
|
args.append("--debug")
|
2019-11-12 14:00:47 +00:00
|
|
|
if self.bootopts:
|
|
|
|
args.extend(("--boot", self.bootopts))
|
2017-02-15 01:41:31 +00:00
|
|
|
if textinst:
|
|
|
|
args.extend(("--graphics", "none", "--extra-args", "console=ttyS0"))
|
|
|
|
else:
|
|
|
|
args.extend(("--graphics", "vnc", "--noautoconsole"))
|
2016-10-21 23:54:03 +00:00
|
|
|
# this is a hacky workaround for a weird bug on Fedora's prod
|
|
|
|
# openQA server:
|
|
|
|
# https://bugzilla.redhat.com/show_bug.cgi?id=1387798
|
|
|
|
args.extend(("--network", "user"))
|
2016-05-23 13:55:17 +00:00
|
|
|
# run the command, timing out after 1 hour; sometimes creation
|
|
|
|
# seems to just get mysteriously stuck, we need to bail and
|
|
|
|
# retry in this case
|
|
|
|
try:
|
2017-02-15 01:41:31 +00:00
|
|
|
logger.info("Install starting...")
|
|
|
|
logger.debug("Command: %s", ' '.join(args))
|
|
|
|
if not textinst:
|
|
|
|
logger.info("Connect via VNC to monitor")
|
2016-05-23 13:55:17 +00:00
|
|
|
ret = subprocess.call(args, timeout=3600)
|
|
|
|
except subprocess.TimeoutExpired:
|
|
|
|
logger.warning("Image creation timed out!")
|
2017-08-31 21:28:04 +00:00
|
|
|
# clean up the domain again
|
|
|
|
conn = libvirt.open()
|
|
|
|
dom = conn.lookupByName('createhdds')
|
|
|
|
try:
|
|
|
|
dom.destroy()
|
|
|
|
except libvirt.libvirtError:
|
|
|
|
# maybe it died already
|
|
|
|
pass
|
|
|
|
try:
|
2018-03-14 20:42:01 +00:00
|
|
|
dom.undefineFlags(libvirt.VIR_DOMAIN_UNDEFINE_NVRAM)
|
2017-08-31 21:28:04 +00:00
|
|
|
except libvirt.libvirtError:
|
|
|
|
pass
|
2017-08-30 21:29:59 +00:00
|
|
|
conn.close()
|
2016-05-23 13:55:17 +00:00
|
|
|
if os.path.isfile(tmpfile):
|
|
|
|
os.remove(tmpfile)
|
|
|
|
if retries:
|
|
|
|
logger.info("Retrying: %s retries remain after this", str(retries))
|
2017-04-07 17:55:20 +00:00
|
|
|
return self.create(textinst, retries=retries-1)
|
2016-05-23 13:55:17 +00:00
|
|
|
else:
|
|
|
|
sys.exit("Image creation timed out too many times!")
|
|
|
|
if ret > 0:
|
|
|
|
# wipe the temp file
|
|
|
|
if os.path.isfile(tmpfile):
|
|
|
|
os.remove(tmpfile)
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
sys.exit("virt-install command {0} failed!".format(' '.join(args)))
|
|
|
|
# at this point the domain should be shut off; if it's
|
|
|
|
# anything else, something went wrong: clean up and exit
|
|
|
|
conn = libvirt.open()
|
|
|
|
dom = conn.lookupByName('createhdds')
|
|
|
|
if dom.state()[0] != libvirt.VIR_DOMAIN_SHUTOFF:
|
|
|
|
conn.close()
|
2016-05-23 13:55:17 +00:00
|
|
|
if os.path.isfile(tmpfile):
|
|
|
|
os.remove(tmpfile)
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
sys.exit("libvirt domain ('createhdds') is not shutdown! "
|
|
|
|
"this is an unexpected condition, aborting.")
|
|
|
|
# we're all done! rename to the correct name and clean up
|
|
|
|
# the domain
|
2016-05-23 13:55:17 +00:00
|
|
|
os.rename(tmpfile, self.filename)
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
os.chmod(self.filename, 0o644)
|
2018-03-14 20:42:01 +00:00
|
|
|
dom.undefineFlags(libvirt.VIR_DOMAIN_UNDEFINE_NVRAM)
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
conn.close()
|
2016-05-23 13:55:17 +00:00
|
|
|
except:
|
|
|
|
# if anything went wrong, we want to wipe the temp file
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
# then raise. we leave the domain intact in case we want
|
|
|
|
# to debug it
|
2016-05-23 13:55:17 +00:00
|
|
|
if os.path.isfile(tmpfile):
|
|
|
|
os.remove(tmpfile)
|
|
|
|
raise
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
|
|
|
@property
|
|
|
|
def outdated(self):
|
|
|
|
"""Whether the image is outdated - if self.maxage is set and
|
|
|
|
it's older than that.
|
|
|
|
"""
|
|
|
|
if not os.path.isfile(self.filename):
|
|
|
|
return False
|
|
|
|
|
|
|
|
if self.maxage:
|
|
|
|
age = int(time.time()) - int(os.path.getmtime(self.filename))
|
|
|
|
# maxage is in days
|
|
|
|
if age > int(self.maxage) * 24 * 60 * 60:
|
|
|
|
return True
|
|
|
|
|
|
|
|
return False
|
|
|
|
|
|
|
|
|
|
|
|
def get_guestfs_images(imggrp, labels=None, filesystems=None):
|
|
|
|
"""Passed a single 'image group' dict (usually read out of hdds.
|
|
|
|
json), returns a list of GuestfsImage instances. labels and
|
|
|
|
filesystems act as overrides to the values specified in hdds.json
|
|
|
|
and the defaults in this function; multiple images will be created
|
|
|
|
for whatever combinations of labels and filesystems are specified.
|
|
|
|
If they're not passed, then we use the values from the dict; if
|
|
|
|
the dict doesn't specify, we just create a single image with the
|
|
|
|
label set to 'mbr' and the filesystem to 'ext4'. The filesystem
|
|
|
|
setting is itself a default - it's only used for 'part' entries
|
|
|
|
which don't specify a filesystem (see the GuestfsImage docs).
|
|
|
|
"""
|
|
|
|
imgs = []
|
|
|
|
# Read in the various dict values
|
|
|
|
name = imggrp['name']
|
|
|
|
size = imggrp['size']
|
|
|
|
parts = imggrp['parts']
|
|
|
|
# These are optional
|
|
|
|
writes = imggrp.get('writes')
|
|
|
|
uploads = imggrp.get('uploads')
|
|
|
|
imgver = imggrp.get('imgver')
|
|
|
|
|
|
|
|
# Here we implement the 'labels / filesystems' behaviour explained
|
|
|
|
# in the docstring
|
|
|
|
if not labels:
|
|
|
|
labels = imggrp.get('labels', ['mbr'])
|
|
|
|
if not filesystems:
|
|
|
|
filesystems = imggrp.get('filesystems', ['ext4'])
|
|
|
|
|
|
|
|
# Now we've sorted out all our settings, instantiate the images
|
|
|
|
for label in labels:
|
|
|
|
for filesystem in filesystems:
|
|
|
|
# We want to indicate filesystem/label in the filename
|
|
|
|
# when we're creating images with more than one. We check
|
|
|
|
# both the passed 'filesystems' value and the imggrp
|
|
|
|
# value, so if the caller overrode the imggrp value we
|
|
|
|
# still get the the long name (happens when using the
|
|
|
|
# single-image CLI subcommand and restricting the label/
|
|
|
|
# filesystem)
|
|
|
|
name_extras = []
|
|
|
|
if len(imggrp.get('filesystems', [])) > 1 or len(filesystems) > 1:
|
|
|
|
name_extras.append(filesystem)
|
|
|
|
if len(imggrp.get('labels', [])) > 1 or len(labels) > 1:
|
|
|
|
name_extras.append(label)
|
|
|
|
img = GuestfsImage(
|
|
|
|
name, size, imgver, filesystem, label, parts, writes, uploads, name_extras)
|
|
|
|
imgs.append(img)
|
|
|
|
return imgs
|
|
|
|
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
def get_virtinstall_images(imggrp, nextrel=None, releases=None):
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
"""Passed a single 'image group' dict (usually read out of hdds.
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
json), returns a list of VirtInstallImage instances. 'nextrel'
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
indicates the 'next' release of Fedora: sometimes we determine the
|
|
|
|
release to build image(s) for in relation to this, so we need to
|
|
|
|
know it. If it's not specified, we ask fedfind to figure it out
|
|
|
|
for us (this is the usual case, we just allow specifying it
|
|
|
|
in case there's an issue with fedfind or you want to test image
|
|
|
|
creation for the next-next release ahead of time or something).
|
|
|
|
The image group dict must include a dict named 'releases' which
|
|
|
|
indicates what releases to build images for and what arches to
|
|
|
|
build for each release; 'releases' can be used to override that. If
|
|
|
|
set, the image group 'releases' dict is ignored, and this dict is
|
|
|
|
used instead. The dict's keys must be release numbers or negative
|
|
|
|
integers: -1 means 'one release lower than the "next" release',
|
|
|
|
-2 means 'two releases lower than the "next" release', and so on.
|
|
|
|
The values are the arches to build for that release.
|
|
|
|
"""
|
2020-05-06 23:04:19 +00:00
|
|
|
imgs = {}
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# Set this here so if we need to calculate it, we only do it once
|
|
|
|
if not nextrel:
|
|
|
|
nextrel = 0
|
|
|
|
name = imggrp['name']
|
|
|
|
# this is the second place we set a default for maxage - bit ugly
|
2022-05-25 21:54:25 +00:00
|
|
|
maxage = int(imggrp.get('maxage', 7))
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
# ditto variant
|
|
|
|
variant = imggrp.get('variant')
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
if not releases:
|
|
|
|
releases = imggrp['releases']
|
|
|
|
size = imggrp.get('size', 0)
|
|
|
|
imgver = imggrp.get('imgver')
|
2019-11-12 14:00:47 +00:00
|
|
|
bootopts = imggrp.get('bootopts')
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# add an image for each release/arch combination
|
2017-02-15 01:41:31 +00:00
|
|
|
for (release, arches) in releases.items():
|
2022-04-26 16:59:23 +00:00
|
|
|
branched = release.lower() == 'branched'
|
|
|
|
rawhide = release.lower() == 'rawhide'
|
|
|
|
if branched:
|
2017-02-15 01:41:31 +00:00
|
|
|
# find Branched, if it exists
|
|
|
|
curr = fedfind.helpers.get_current_release(branched=False)
|
|
|
|
branch = fedfind.helpers.get_current_release(branched=True)
|
|
|
|
if branch > curr:
|
2017-07-27 19:56:26 +00:00
|
|
|
rels = [branch]
|
2017-02-15 01:41:31 +00:00
|
|
|
else:
|
|
|
|
logger.info("Branched image requested, but Branched does not currently exist")
|
|
|
|
continue
|
2022-04-26 16:59:23 +00:00
|
|
|
elif rawhide:
|
|
|
|
# find Rawhide release number
|
|
|
|
rawrel = fedfind.helpers.get_current_release(branched=True) + 1
|
|
|
|
rels = [rawrel]
|
2017-07-27 19:56:26 +00:00
|
|
|
elif release.lower() == 'stable':
|
|
|
|
# this means "all current stable releases"
|
|
|
|
rels = fedfind.helpers.get_current_stables()
|
2017-02-15 01:41:31 +00:00
|
|
|
elif release != 'rawhide' and int(release) < 0:
|
|
|
|
# negative release indicates 'relative to next release'
|
2021-02-19 19:29:25 +00:00
|
|
|
# -1 is CURRREL, -2 is PREVREL
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
if not nextrel:
|
|
|
|
nextrel = fedfind.helpers.get_current_release() + 1
|
2017-07-27 19:56:26 +00:00
|
|
|
rels = [int(nextrel) + int(release)]
|
|
|
|
else:
|
2022-04-26 16:59:23 +00:00
|
|
|
# assume a single integer release number
|
2017-07-27 19:56:26 +00:00
|
|
|
rels = [release]
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
for arch in arches:
|
2017-07-27 19:56:26 +00:00
|
|
|
for rel in rels:
|
2020-05-06 23:04:19 +00:00
|
|
|
key = "{0}-{1}".format(rel, arch)
|
|
|
|
# using a dict here avoids dupes
|
|
|
|
imgs[key] = VirtInstallImage(name, rel, arch, variant=variant, size=size,
|
2022-04-26 16:59:23 +00:00
|
|
|
imgver=imgver, maxage=maxage, bootopts=bootopts,
|
|
|
|
branched=branched, rawhide=rawhide)
|
2020-05-06 23:04:19 +00:00
|
|
|
return list(imgs.values())
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
|
|
|
def get_all_images(hdds, nextrel=None):
|
|
|
|
"""Simply iterates over the 'image group' dicts in hdds.json and
|
|
|
|
calls the appropriate get_foo_images() function for each, then
|
|
|
|
returns the list of all images. No overrides for labels,
|
|
|
|
filesystems, releases or arches are provided here, this function
|
|
|
|
is just for painting inside the lines - it's used to determine
|
|
|
|
what images are 'expected' to exist according to hdds.json. We
|
|
|
|
do allow passing of 'nextrel' just in case there's some issue
|
|
|
|
with the auto-discovery.
|
|
|
|
"""
|
|
|
|
imgs = []
|
|
|
|
for imggrp in hdds['guestfs']:
|
|
|
|
imgs.extend(get_guestfs_images(imggrp))
|
|
|
|
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
for imggrp in hdds['virtinstall']:
|
|
|
|
imgs.extend(get_virtinstall_images(imggrp, nextrel=nextrel))
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
return imgs
|
|
|
|
|
|
|
|
def do_renames(hdds):
|
|
|
|
"""Rename files according to the 'renames' list in hdds.json,
|
|
|
|
which is just a list of 'old name, new name' pairs. Say there's
|
|
|
|
an image we used to only create an mbr version of, but now we add
|
|
|
|
a gpt version; that means the expected name of the mbr version
|
|
|
|
changes. This function allows us to specify in hdds.json that the
|
|
|
|
mbr image should be renamed, which saves having to rebuild it.
|
|
|
|
'all' always does renames, 'check' can be told to do renames.
|
|
|
|
"""
|
|
|
|
for (orig, new) in hdds['renames']:
|
|
|
|
if os.path.isfile(orig) and not os.path.exists(new):
|
|
|
|
logger.info("Renaming %s to %s...", orig, new)
|
|
|
|
os.rename(orig, new)
|
|
|
|
|
|
|
|
def delete_all():
|
|
|
|
"""Remove absolutely all createhdds-controlled files; we assume
|
|
|
|
anything in the working directory starting with 'disk' and ending
|
|
|
|
with 'img' is one of our files.
|
|
|
|
"""
|
|
|
|
files = [fl for fl in os.listdir('.') if fl.startswith('disk') and fl.endswith('img')]
|
|
|
|
for _file in files:
|
|
|
|
os.remove(_file)
|
|
|
|
|
|
|
|
def clean(unknown):
|
|
|
|
"""This simply removes all the files in the list. The list is
|
|
|
|
expected to be the list of 'unknown' files returned by check();
|
|
|
|
both 'all' and 'clean' can delete 'unknown' files if requested.
|
|
|
|
'unknown' files are usually images for old releases we're no
|
|
|
|
longer testing, images we've simply stopped using, or old image
|
|
|
|
files when the image group name has been changed to indicate an
|
|
|
|
incompatible change to the image(s).
|
|
|
|
"""
|
|
|
|
for filename in unknown:
|
|
|
|
try:
|
|
|
|
logger.info("Removing unknown image %s", filename)
|
|
|
|
os.remove(filename)
|
|
|
|
except OSError:
|
|
|
|
# We don't really care if the file didn't exist for some
|
|
|
|
# reason, so just pass.
|
|
|
|
pass
|
|
|
|
|
|
|
|
def check(hdds, nextrel=None):
|
|
|
|
"""This calls get_all_images() to find out what images are expected
|
|
|
|
to exist, then compares that to the images that are actually
|
|
|
|
present and returns three lists. The first two are lists of Image
|
|
|
|
subclass instances, the last is a list of filenames. The first is a
|
|
|
|
list of the images that just aren't there at all. The second is a
|
|
|
|
list of images that are present but 'outdated', because they've
|
|
|
|
exceeded their maxage. The last is a list of image files that are
|
|
|
|
present but don't match any of the expected images - 'unknown'
|
|
|
|
images. The 'clean()' function can be used to remove these.
|
|
|
|
'nextrel' is passed through to get_all_images(), and is available
|
|
|
|
in case auto-detection via fedfind fails.
|
|
|
|
"""
|
|
|
|
current = []
|
|
|
|
missing = []
|
|
|
|
outdated = []
|
|
|
|
unknown = []
|
|
|
|
# Get the list of all 'expected' images
|
|
|
|
expected = get_all_images(hdds, nextrel=nextrel)
|
|
|
|
|
|
|
|
# Now determine if images are absent or outdated
|
|
|
|
for img in expected:
|
2017-10-16 16:51:56 +00:00
|
|
|
if img.arch in supported_arches() and not os.path.isfile(img.filename):
|
2021-09-10 17:24:40 +00:00
|
|
|
# handle renaming qcow2 images from old ('.img') to new ('.qcow2')
|
|
|
|
oldfn = img.filename.replace(".qcow2", ".img")
|
|
|
|
if os.path.isfile(oldfn):
|
2021-09-10 18:00:32 +00:00
|
|
|
os.rename(oldfn, img.filename)
|
2021-09-10 17:24:40 +00:00
|
|
|
if img.outdated:
|
|
|
|
outdated.append(img)
|
|
|
|
else:
|
|
|
|
missing.append(img)
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
continue
|
|
|
|
if img.outdated:
|
|
|
|
outdated.append(img)
|
|
|
|
continue
|
|
|
|
current.append(img)
|
|
|
|
|
2021-09-10 18:07:58 +00:00
|
|
|
# Get the list of all present image files
|
|
|
|
files = set(fl for fl in os.listdir('.') if fl.startswith('disk') and (fl.endswith('img') or fl.endswith('qcow2')))
|
|
|
|
# Compare present images vs. expected images to produce 'unknown'
|
|
|
|
expnames = set(img.filename for img in expected)
|
|
|
|
unknown = list(files.difference(expnames))
|
|
|
|
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
logger.debug("Current images: %s", ', '.join([img.filename for img in current]))
|
|
|
|
logger.debug("Missing images: %s", ', '.join([img.filename for img in missing]))
|
|
|
|
logger.debug("Outdated images: %s", ', '.join([img.filename for img in outdated]))
|
|
|
|
logger.debug("Unknown images: %s", ', '.join(unknown))
|
|
|
|
return (missing, outdated, unknown)
|
|
|
|
|
|
|
|
def cli_all(args, hdds):
|
|
|
|
"""Function for the CLI 'all' subcommand. Creates all images. If
|
|
|
|
args.delete is set, blows all existing images away and recreates
|
|
|
|
them; otherwise it will just fill in missing or outdated images.
|
|
|
|
If args.clean is set, also wipes 'unknown' images.
|
|
|
|
"""
|
|
|
|
if args.delete:
|
|
|
|
logger.info("Removing all images...")
|
|
|
|
delete_all()
|
|
|
|
|
|
|
|
# handle renamed images (see do_renames docstring)
|
|
|
|
do_renames(hdds)
|
|
|
|
|
|
|
|
# call check() to find out what we need to do
|
|
|
|
(missing, outdated, unknown) = check(hdds, nextrel=args.nextrel)
|
|
|
|
|
|
|
|
# wipe 'unknown' images if requested
|
|
|
|
if args.clean:
|
|
|
|
clean(unknown)
|
|
|
|
|
|
|
|
# 'missing' plus 'outdated' is all the images we need to build; if
|
|
|
|
# args.delete was set, all images will be in this list
|
|
|
|
missing.extend(outdated)
|
|
|
|
for (num, img) in enumerate(missing, 1):
|
|
|
|
logger.info("Creating image %s...[%s/%s]", img.filename, str(num), str(len(missing)))
|
2017-02-15 01:41:31 +00:00
|
|
|
img.create(args.textinst)
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
|
|
|
def cli_check(args, hdds):
|
|
|
|
"""Function for the CLI 'check' subcommand. Basically just calls
|
|
|
|
check() and prints the results. Does renames before checking if
|
|
|
|
args.rename was set, and wipes 'unknown' images after checking if
|
2016-10-17 18:26:18 +00:00
|
|
|
args.clean was set. Exits with status 2 if any images are missing,
|
|
|
|
1 if all images are present but one or more is outdated.
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
"""
|
|
|
|
if args.rename:
|
|
|
|
do_renames(hdds)
|
|
|
|
|
|
|
|
(missing, outdated, unknown) = check(hdds, nextrel=args.nextrel)
|
|
|
|
if missing:
|
|
|
|
print("Missing images: {0}".format(', '.join([img.filename for img in missing])))
|
|
|
|
if outdated:
|
|
|
|
print("Outdated images: {0}".format(', '.join([img.filename for img in outdated])))
|
|
|
|
if unknown:
|
|
|
|
print("Unknown images: {0}".format(', '.join(unknown)))
|
|
|
|
if args.clean:
|
|
|
|
clean(unknown)
|
|
|
|
|
2016-10-17 18:26:18 +00:00
|
|
|
if missing:
|
|
|
|
sys.exit(2)
|
|
|
|
elif outdated:
|
|
|
|
sys.exit(1)
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
else:
|
|
|
|
sys.exit()
|
|
|
|
|
|
|
|
def cli_image(args, *_):
|
|
|
|
"""Function for CLI image group subcommands (a subcommand is added
|
|
|
|
for each image group in hdds.json). Will create the image(s) from
|
|
|
|
the specified group. For guestfs image groups with multiple labels
|
|
|
|
and/or filesystems, the user can pass args.label and/or args.
|
|
|
|
filesystem to limit creation to a single label and/or filesystem.
|
|
|
|
Note this function does no checking; the image will always be
|
|
|
|
recreated, even if it already exists and is current.
|
|
|
|
"""
|
|
|
|
# Note that on this path, the parsing of hdds is done by
|
|
|
|
# parse_args(). It passes us the image type and the image group
|
|
|
|
# dict from hdds as a tuple; we need to know the type so we know
|
|
|
|
# what fiddling to do with the args and what function to call
|
|
|
|
imggrp = args.imggrp[1]
|
|
|
|
imgtype = args.imggrp[0]
|
|
|
|
|
|
|
|
if imgtype == 'guestfs':
|
|
|
|
# If the user passed in label or filesystem, we pass them on
|
|
|
|
# to get_guestfs_images as single-item lists, otherwise we
|
|
|
|
# just pass 'None', causing it to use the values from imggrp
|
|
|
|
# (which come from hdds.json)
|
|
|
|
labels = None
|
|
|
|
filesystems = None
|
|
|
|
if args.label:
|
|
|
|
labels = [args.label]
|
|
|
|
if args.filesystem:
|
|
|
|
filesystems = [args.filesystem]
|
|
|
|
imgs = get_guestfs_images(imggrp, labels=labels, filesystems=filesystems)
|
|
|
|
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
elif imgtype == 'virtinstall':
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# If the user passed args.release, we construct a releases
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
# dict to pass to get_virtinstall_images to override the dict
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# from imggrp. If they passed args.arch, we use that arch,
|
|
|
|
# otherwise we default to x86_64. If args.release isn't set,
|
|
|
|
# we just pass None as release, and the releases dict from
|
|
|
|
# imggrp will be used, and images created for all release/
|
|
|
|
# arch combinations listed there. FIXME: if the user passes
|
|
|
|
# args.arch but not args.release, we just ignore it...
|
|
|
|
releases = None
|
|
|
|
if args.release:
|
|
|
|
if args.arch:
|
|
|
|
arches = [args.arch]
|
|
|
|
else:
|
|
|
|
arches = ['x86_64']
|
|
|
|
releases = {args.release: arches}
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
imgs = get_virtinstall_images(imggrp, releases=releases)
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
|
|
|
for (num, img) in enumerate(imgs, 1):
|
|
|
|
logger.info("Creating image %s...[%s/%s]", img.filename, str(num), str(len(imgs)))
|
2017-02-15 01:41:31 +00:00
|
|
|
img.create(args.textinst)
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
|
|
|
def parse_args(hdds):
|
|
|
|
"""Parse arguments with argparse."""
|
|
|
|
parser = argparse.ArgumentParser(description=(
|
|
|
|
"Tool for creating hard disk images for Fedora openQA."))
|
|
|
|
parser.add_argument(
|
|
|
|
'-l', '--loglevel', help="The level of log messages to show",
|
|
|
|
choices=('debug', 'info', 'warning', 'error', 'critical'),
|
|
|
|
default='info')
|
2017-02-15 01:41:31 +00:00
|
|
|
parser.add_argument(
|
|
|
|
'-t', '--textinst', help="For any virt-install images, run the install in text mode "
|
|
|
|
"and show details on stdout", action='store_true')
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
|
|
|
|
# This is a workaround for a somewhat infamous argparse bug
|
|
|
|
# in Python 3. See:
|
|
|
|
# https://stackoverflow.com/questions/23349349/argparse-with-required-subparser
|
|
|
|
# http://bugs.python.org/issue16308
|
|
|
|
subparsers = parser.add_subparsers(dest='subcommand')
|
|
|
|
subparsers.required = True
|
|
|
|
|
|
|
|
parser_all = subparsers.add_parser(
|
|
|
|
'all', description="Ensure all images are present and up-to-date.")
|
|
|
|
parser_all.add_argument(
|
|
|
|
'-d', '--delete', help="Delete and re-build all images",
|
|
|
|
action='store_true')
|
|
|
|
parser_all.add_argument(
|
|
|
|
'-c', '--clean', help="Remove unknown (usually old) images",
|
|
|
|
action='store_true')
|
|
|
|
parser_all.add_argument(
|
|
|
|
'-n', '--nextrel', help="The release to treat as the 'next' release "
|
|
|
|
"- this determines what releases some images will be built for. If "
|
|
|
|
"not set or set to 0, createhdds will try to discover it when needed",
|
|
|
|
type=int, default=0)
|
|
|
|
parser_all.set_defaults(func=cli_all)
|
|
|
|
|
|
|
|
parser_check = subparsers.add_parser(
|
|
|
|
'check', description="Check status of existing image files.")
|
|
|
|
parser_check.add_argument(
|
|
|
|
'-n', '--nextrel', help="The release to treat as the 'next' release "
|
|
|
|
"- this determines what releases some images are expected to exist for. If "
|
|
|
|
"not set or set to 0, createhdds will try to discover it when needed",
|
|
|
|
type=int, default=0)
|
|
|
|
parser_check.add_argument(
|
|
|
|
'-r', '--rename', help="Whether to rename images when the expected name "
|
|
|
|
"has changed or not", action="store_true")
|
|
|
|
parser_check.add_argument(
|
|
|
|
'-c', '--clean', help="Remove unknown (usually old) images",
|
|
|
|
action='store_true')
|
|
|
|
parser_check.set_defaults(func=cli_check)
|
|
|
|
|
|
|
|
# This here is somewhat clever-clever: we generate a subcommand for
|
|
|
|
# each image group listed in hdds.json, which can be used to build
|
|
|
|
# image(s) from that group. For guestfs image groups, we also check
|
|
|
|
# if the group has multiple labels and/or filesystems by default,
|
|
|
|
# and add arguments to limit image creation to just a single label
|
|
|
|
# and/or filesystem.
|
|
|
|
for imggrp in hdds['guestfs']:
|
|
|
|
imgparser = subparsers.add_parser(
|
|
|
|
imggrp['name'], description="Create {0} image(s)".format(imggrp['name']))
|
|
|
|
labels = imggrp.get('labels', [])
|
|
|
|
if len(labels) > 1:
|
|
|
|
imgparser.add_argument(
|
|
|
|
'-l', '--label', help="Create only images with this disk label",
|
|
|
|
choices=labels)
|
|
|
|
filesystems = imggrp.get('filesystems', [])
|
|
|
|
if len(filesystems) > 1:
|
|
|
|
imgparser.add_argument(
|
|
|
|
'-f', '--filesystem', help="Create only images with this filesystem",
|
|
|
|
choices=filesystems)
|
|
|
|
imgparser.set_defaults(func=cli_image, label=None, filesystem=None)
|
|
|
|
# Here we're stuffing the type of the image and the dict from
|
|
|
|
# hdds into args for cli_image() to use.
|
|
|
|
imgparser.set_defaults(imggrp=('guestfs', imggrp))
|
|
|
|
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
# For virtinstall images, we provide args to override the release/
|
|
|
|
# arch combination; using args.release will always result in just a
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
# single image being built, for x86_64 unless args.arch is set to
|
|
|
|
# i686.
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
for imggrp in hdds['virtinstall']:
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
imgparser = subparsers.add_parser(
|
|
|
|
imggrp['name'], description="Create {0} image(s)".format(imggrp['name']))
|
|
|
|
imgparser.add_argument(
|
|
|
|
'-r', '--release', help="The release to build the image(s) for. If not "
|
2017-02-15 01:41:31 +00:00
|
|
|
"set, createhdds will attempt to determine the current release and build "
|
|
|
|
"for appropriate releases relative to that",
|
|
|
|
default='')
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
imgparser.add_argument(
|
|
|
|
'-a', '--arch', help="The arch to build the image(s) for. If neither "
|
|
|
|
"this nor --release is set, createhdds will decide the appropriate "
|
|
|
|
"arch(es) to build for each release. If this is not set but --release "
|
|
|
|
"is set, only x86_64 image(s) will be built.",
|
2020-04-18 22:15:04 +00:00
|
|
|
choices=('x86_64', 'i686', 'ppc64le', 'aarch64'))
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
imgparser.set_defaults(func=cli_image)
|
|
|
|
# Here we're stuffing the type of the image and the dict from
|
|
|
|
# hdds into args for cli_image() to use.
|
switch from virt-builder to virt-install (T813)
Summary:
We've kinda been having too much trouble with virt-builder
lately, mainly SELinux related issues due to how it does image
customization. It also produces images that differ in notable
ways from what a 'typical' install would give. virt-install
solves both these problems, and also gives us more flexibility
for storage configuration and post-install customization should
we need them in future.
The change isn't really too drastic, and the design is similar:
instead of virt-builder commands files, each image type now has
a kickstart file where all its customizations can be done.
There's also a single extra image dict key, 'variant', which
specifies which install tree variant to use for running the
install. It defaults to 'Everything' (for F24+) and 'Server'
(for <F24, as Everything wasn't installable until F24) but we
set it to 'Server' for the server images and 'Workstation' for
the desktop images, so those installs will use the correct
variant install class.
We run the installs in VNC. You can do it with a serial console
and log the output, but then anaconda gets clever and changes
several things in the installed system based on the fact that
you did the install over a serial console: it twiddles with
the kernel args and doesn't set graphical.target as the default.
We don't want any of that mess, so we do a VNC install.
The 'size' value is just a number of gigabytes for virt-install
images (as that's how the virt-install 'size' argument works).
This also drops some unused 32-bit images (we don't do 32-bit
KDE or Server upgrade tests, so there's no need to build those
images).
Test Plan:
Re-generate all affected images and re-run all tests
that use them, make sure they work. I am doing this on staging
at present. Note: this would render D911 unnecessary.
Reviewers: garretraziel
Reviewed By: garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D917
2016-07-04 16:29:25 +00:00
|
|
|
imgparser.set_defaults(imggrp=('virtinstall', imggrp))
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
return parser.parse_args()
|
|
|
|
|
|
|
|
def main():
|
|
|
|
"""Main loop - set up logging, parse args, run subcommand
|
|
|
|
function.
|
|
|
|
"""
|
|
|
|
try:
|
|
|
|
with open('{0}/hdds.json'.format(SCRIPTDIR), 'r') as fout:
|
|
|
|
hdds = json.load(fout)
|
|
|
|
args = parse_args(hdds)
|
|
|
|
loglevel = getattr(
|
|
|
|
logging, args.loglevel.upper(), logging.INFO)
|
|
|
|
logging.basicConfig(level=loglevel)
|
|
|
|
args.func(args, hdds)
|
|
|
|
except KeyboardInterrupt:
|
|
|
|
sys.stderr.write("Interrupted, exiting...\n")
|
2016-05-23 13:55:17 +00:00
|
|
|
# there may be temp image files lying around...
|
rewrite createhdds in python, make it shinier
Summary:
createhdds.sh was just too damn simple and understandable, so
I thought I'd make it three times longer, object oriented,
and hard to understand!
OK, OK, that's not great sales. Alright. The main thing was to
make it smarter. This rewrite lets it do these things:
* Only create the images that are missing (not rebuild all)
* Work out the releases to build images for
* Rename images when appropriate
* Rebuild images when they need rebuilding
* Remove old / abandoned images
It can figure out what images ought to be present - including
working out the 'next' release and figuring out from that what
releases it needs images for - and build only the missing ones.
There's a 'version' concept for images; if the existing image
is older than the version given in the data file, it'll be
rebuilt. The data file can list 'rename' pairs, allowing images
to be renamed (like when we move from a single image to multiple
label/filesystem variants). This code uses fedfind's ability
to find the current release version to figure out what releases
we need virtbuilder images for (so you don't have to pass it
in). And it can find image files that aren't in the 'currently
expected' set and wipe them. Images can also have a 'maxage',
triggering a rebuild when they exceed it - this is intended
for the virtbuilder images, so we get a rebuild with the
latest updates every so often (default is two weeks).
The point of all this is to help with unattended deployment/
maintenance, i.e. the ansible deployment we have in infra;
the idea is that we can just set that up to run the 'all'
subcommand every so often, and it'll remove old images, create
new ones, and rebuild ones that are outdated.
I kept the ability to build a single image (or a whole image
'group'), and included the ability to just run a check without
actually doing a rebuild. There's a few little weird things
and holes here as it's not really the focus of the tool.
Test Plan:
Build all images, do a full test run, and see if
it works OK. Test out all the variations of building single
images / image groups, and using the 'check' command.
Reviewers: jskladan, garretraziel
Reviewed By: jskladan, garretraziel
Subscribers: tflink
Differential Revision: https://phab.qadevel.cloud.fedoraproject.org/D687
2015-12-17 21:25:30 +00:00
|
|
|
tmps = [fl for fl in os.listdir('.') if fl.startswith('disk') and fl.endswith('.tmp')]
|
|
|
|
for tmp in tmps:
|
|
|
|
os.remove(tmp)
|
|
|
|
sys.exit(1)
|
|
|
|
|
|
|
|
if __name__ == '__main__':
|
|
|
|
main()
|