When the repository has multiple arches, eg. i686 and x86_64, it should
add a new entry to the project's builds list, not create a new project
in the list.
This handles that by adding a modified insort_left function and
examining the packages returned from dnf to make sure they aren't
already listed in the results. It also handles adding them in sorted
order so that no further sorting needs to be done on the results.
Resolves: rhbz#1657055
(cherry picked from commit 663a0dcd73)
If the system ran out of space, or was rebooted unexpectedly, the state
of the queue symlinks, or the results STATUS files may be inconsistent.
This checks them and:
* Removes broken symlinks from queue/new and queue/run
* Removes symlinks from run and sets the build to FAILED
* Sets builds w/o a STATUS to FAILED
* Sets builds with STATUS of RUNNING to FAILED
* Creates missing queue/new symlinks to results with STATUS of WAITING
So, any builds that were running during the reboot will be FAILED, and
any that were waiting to be started will be started upon rebooting.
Resolves: rhbz#1657054
(cherry picked from commit f0bac40d7f)
Since these images can be used to create multiple machines, they should
not have a unique machine-id attached to them. Replace /etc/machine-id
with an empty file so that it will be regenerated at boot time.
(cherry picked from commit 6fab72d894)
Related: rhbz#1656105
This is similar to the AMI type, but also adds open-vm-tools and does not do
anything special to the partitioning
(cherry picked from commit 1056bfc25b)
Resolves: rhbz#1656105
This is similar to the AMI compose type, with a handful of additional
changes specific to Azure:
* Add waagent (but leave NetworkManager enabled, despite some of the
docs)
* Disable cloud-init
* Add Hyper-V modules into initrams.
Fixes specific for RHEL:
* Create ifcfg-eth0 required by waagent.
* Install python3 and net-tools required by waagent.
Recommended changes:
* Use recommended kernel boot args.
* Disable kdump.
(cherry picked from commit e0c236ff36)
(cherry picked from commit da0435bc90)
(cherry picked from commit b594fa99bc)
Resolves: rhbz#1656105
This differs from lmc's --make-ami in that creates a full disk image instead of
an fsimage. Create a raw disk image with a / and /boot partitions, and enable
sshd, chronyd, and cockpit by default.
(cherry picked from commit 18188bf6cf)
(cherry picked from commit 81d38b6445)
Resolves: rhbz#1656105
This is required to ensure that SELinux is configured properly while
building. It fixes the problem with building tar, and should be
installed in the other image types for consistency.
Resolves: rhbz#1654795
Anaconda, Lorax, lorax-composer, and livemedia-creator can all now run
with SELinux in Enforcing mode. It does not need to be disabled and if
there are denials they should be reported as a bug.
Log the current state of SELinux when starting, update the
documentation.
Resolves: rhbz#1654795
Some kickstart templates use globbing to match packages, searchNames
doesn't support this.
Add fallback code to find packages that are just a dep, matching the
behavior of yum.install()
This fixes things so that it can depsolve package names with globs in
them like grub2-efi-*-cdboot, and deps like shim and grub2-efi.
Results are still filtered by the version glob, with the highest NEVRA
selected for installation.
Resolves: rhbz#1641601
This tests to make sure that the metadata timer is working (by setting
it to 10s and adding a new package to the repo), and that
YumLock.lock_check immediately picks up a new package.
This depends on rpmfluff which is available from Fedora or EPEL repos.
Related: rhbz#1632962
The problem this solves is that yum really isn't designed to be part of\
a long running daemon. So when repodata changes upstream, even when
you force it to download the new metadata, it doesn't change in memory
so you end up with lorax-composer depsolving against old versions, and
anaconda depsolving against new versions (because it sets up its own
YumBase and cache) and then the kickstart is no longer valid.
To solve this I have
- Added a 6h timeout to the metadata check (because yum's doesn't work
in this situation).
- Added a metadata check to the YumLock .lock property, but only when
the timeout expires.
- Added a new .lock_check property to YumLock that always checks the
metadata and resets the timeout.
If it has changed it does its best to tear down the existing YumBase,
deleting as much as it can in hopes it doesn't leak memory. And then it
sets up a totally new YumBase with the new repodata.
Resolves: rhbz#1632962
Use a common _depsolve function for projects_depsolve and
projects_depsolve_with_size so that it always uses the correct version
glob support when depsolving blueprints and templates.
Resolves: rhbz#1628114
Ends up you cannot use the kickstart user command on root, since it
already exists, so we have to translate that into a rootpw command.
So [[customizations.user]] with name = "root" only support key, which
will set the ssh key, and password which will use rootpw to set the
password. plain text or encrypted are supported.
Related: rhbz#1626120
If we leave the root account w/o a password people will use it that way,
leading to insecure images. Also if we use a default password. So lock
the root account in the templates.
Users will need to do one of these things:
1. Use [[customizations.user]] in their blueprint to configure root or
another user.
2. Use [[customizations.sshkey]] to set a key for root
2. Install a package that configures a user at install time
3. Install a package that sets up a user at boot time (eg. cloud-init)
This also drops the auth line from the kickstart templates, allowing it
to use the default password algoritm instead of md5.
Resolves: rhbz#1626120
The blueprint version glob was being applied to the whole package NEVRA
by yum (it lacks a separate API for just globbing versions), so this
implements that in filterVersionGlob using fnmatchcase on the package
names, and the yum package verGT comparison on the versions for the
selected package.
Also includes tests.
Resolves: rhbz#1628114
blueprints/changes is different, each blueprint has it's own total,
limited by the call's limit. So it needs to find the max total of all
the requested blueprints.
Passing ?limit=0 to the blueprints/list, blueprints/changes,
projects/list, modules/list should always return the total possible
results, not 0.
Also move the composer-cli test_diff to the end so that it will work
consistently. Do this by naming it test_z_diff.
The blueprints/changes API is a bit different from the others, the total
that it includes is for each blueprint, not one total for all of them,
since there will be a different number of commits for each.
The function is passed the dict, and it can be used to select the total
to use for retrieving all of the results. If it isn't included it will
use data["total"] which works fine in most cases.
A value of 1 is too low for heavy users of the API, such as the weldr-web
interface.
This is also systemd's default for sockets it opens. Using lorax-composer with
socket activation already results in a backlog of SOMAXCONN connections.
(cherry picked from commit be5d50e6f3)
Update the docs to make it clear that .socket will be slow to respond to
the first request, and to use .service to speed it up.
(cherry picked from commit 83e5f3f885)
(cherry picked from commit c2f96b3d3b)
Add a limit argument to all potentially paginated results, equal to
whatever the composer backend is the total number of results. This still
has the potential to provide truncated data if the number of results
increases between the two HTTP requests.
Resolves: #404
This adds the following optional arguments to the /compose/status route:
- type, matches the compose_type field
- status, matches the queue_status field
- blueprint, matches the blueprint field
Currently we are making MBR disk images for qcow2 and partitioned disk,
so the UEFI packages aren't required at this point.
Move the clearpart command into compose.py so that in the futute it can
use clearpart --disklabel to create a GPT image, and add the required
packages to the package set.
