This commit refactors VolumeManager to turn it into a proper
factory class and to also include type hints to facilitate it's
use from an API POV. Related to #1498
This commit refactors DiskFormat to turn it into a proper
factory class and to also include type hints to facilitate it's
use from an API POV. Related to #1498
This changes the s390 support on several stages:
1) On s390 the boot process is based on zipl which boots into an
initrd from which a userspace grub process is started to support
the grub capabilities. The implementation of this concept is
provided via the grub2-s390x-emu package. Once installed the
setup of the bootloader is done via the grub2-mkconfig and
grub2-install commands and therefore from a caller perspective
the same as with any other grub2 setup process. For kiwi this
means no extra zipl bootloader target code is needed. Therefore
this commit deletes the zipl setup from kiwi and puts on
the standard grub2 process. This Fixes bsc#1170863
2) To support different targettypes the grub2-s390x-emu provided
zipl template must be adapted. Parts of the former zipl bootloader
setup therefore now applies to an update of the zipl2grub
template file
3) Support for CDL/LDL DASD targets has been disabled in the schema
When testing 4k devices and a respective zipl2grub template
setup for CDL/LDL targettype it has turned out that grub2-install
is not able to run on such a device. My assumption is that
the device code in grub2-install does not work for 4k devices
with an fdasd created partition table. As this needs further
investigations and most probably adaptions on the grub toolchain
for s390, we disabled the setup of these modes for now.
emulated DASD (FBA) and SCSI targets stays supported.
Instead of placing the file into .coveragerc use the
no cover pragma to skip code paths that can never be
reached from being counted in the coverage report.
This affects the abstract constructors in the Factory
classes
This commit refactors BootLoaderInstall class to make it a proper
factory class. In addition type hints are added for the constructor
method.
Related to #1498
This commit refactors BootLoaderConfig to turn it into a proper factory
class and to also include type hints to facilitate it's use from an API
POV.
Related to #1498
In addition to the .packages file which shows details about
the installed packages in terms of version, license, etc...
we now also create a .changes file which contains the changelog
information of the installed packages. The file can be used
to compare the package changelogs between image builds.
This commit refactors the BootImage factory to be a real
factory and to add type hints such that its use from an api
perspective is clear and enforced. Related to Issue #1498
This commit refactors the ImageBuilder factory to be a real
factory and to add type hints such that its use from an api
perspective is clear and enforced. Related to Issue #1498
A vmx image is the same disk as an oem just without the dracut
repart/resize feature. This difference is better handled with
an oemconfig parameter <oem-resize> which allows to switch resize
on or off. The extra image type vmx will be dropped and an XSLT
stylesheet automatically transforms a vmx description to be a
oem image with the resize feature switched off.
This Fixes#1425
This commit removes the creation of the md5 file that includes a
checksum of the image binary for the `tbz` image type. Removing it
as this is the only image type that includes it as part of the result
and because the bundle procedure already creates a sha256 file out of
the results, so there is still the chance to produce validation
checksums.
With regards to Issue #1486 a discussion came up that the
way factories are implemented are questionable when thinking
about strong typing for the public kiwi interface. This
commit refactors the FileSystem factory to be a real factory
and to add type hints such that its use from an api perspective
is clear and enforced.
Allow to put a disk.sh script as part of the image description
which is called for the disk image types `vmx` and `oem`
only and runs after the synchronisation of the root tree into the
disk image loop file. At call time of the script the device name
of the currently mapped root device is passed as a parameter.
The chroot environment for the call is the mounted disk itself
which makes this different from the config.sh/images.sh caller
environment. This Fixes#1464
In preparation to allow a chroot operation into the loop
mounted disk this commit refactors the process when filesystems
gets umounted and also fixes the canonical order for calling
the destructors. Related to Issue #1464
Passing the dmsquash flag adds dmsquash-live and livenet modules to
dracut. This broke when adding a check to only add modules that dracut
reports as available. Use a list instead of a string to represent the
modules to add.
Fixes: 07ea23a4
KIS is an abbreviation for Kernel, Initrd, System and defines a
highly customizable image consisting out of these components.
This commit performs the changes documented in #1414 and
introduces the new kis type. From an image build perspective
kis is currently the same as pxe with restrictions for kis
on the schema level. A kis build uses dracut and does not allow
to use the legacy netboot initrd. The pxe type will therefore be
exclusively used to built for the legacy netboot infrastructure
and is on its way to deprecation in the future.
This Fixes#1262
The new element controls the behavior of the repart/resize code
in the oem-repart dracut module. By default the repart/resize
happens on every reboot and therefore also allows for disk
geometry changes during the livetime of the machine. If the
element is set to false the repart/resize operation happens
only once and then never again. To check for this condition
a new profile environment variable kiwi_rootpartuuid which
holds the PARTUUID of the root partition has been added to
the disk builder.
With the new Fstab class from prior pull request there is an
opportunity to handle all fstab related actions to be done
by that class. This commit extends the Fstab class with an
add_entry method such that we can avoid the extra lists
holding raw fstab lines in e.g the disk builder. In the end
all fstab related data is stored in an instance of the Fstab
class. This also extends the KIWI api by an fstab management
class. Related to #1329 and #1349
Any mount point directly under / should be just right after the root
mountpoint and before the custom mountpoints based on user's subvolume
configuration.
Fixes#1349 and bsc#1164310
The dracut code in 90kiwi-dump/kiwi-dump-image.sh looks for a
file matching ${image_uri}.config.bootoptions but the install
code packs a file named pxeboot.${image_uri}.config.bootoptions
into the tarball. Thus without renaming the file it won't
be found. Also the documentation mentions the file to be
named ${image_uri}.config.bootoptions. This commit fixes the
install code to match the dracut boot code and the documentation
This commit adds the `squashfscompression` attribute in type element. It
can take `gzip`, `zstd`, `xz`, `lzo`, `lz4` or `none`. The default is `xz`.
Fixes#1315
This commit creates the .append file for PXE images. This file
is only created for initrd based images and includes the
UUID of the rootfs and any parameter added by the user within
the 'kernelcmdline' attribute.
Fixes#1327
When deploying a disk image via PXE the initrd contained a config
file which connects it to a certain image. This has the disadvantage
that no other image could be deployed with it. This commit changes the
deployment code in a way that the config file is read from the
network if the disk is deployed via PXE. The tarball created by
kiwi provides the image connected config file but users now have
the opportunity to create their own boot configurations which allows
deployment of different images with the same kiwi built deployment
initrd. This Fixes#1298 and is one first step into a more generic
PXE support offered by kiwi.
If a volume manager is used the volumes are added before the
root filesystem(/) entry in fstab. This does not hurt because
at boot time systemd manages the mounting of the rootfs prior
to any other information in the fstab file but it's conceptually
broken. Users justifiably can expect the fstab entries in the
correct order such that mounting from top to bottom leads
to a consistent root filesystem state.
Added new type attribute:
```xml
<type ... spare_part_fs_attributes="..."/>
```
which can be a comma separated list of the following currently
supported filesystem attributes:
* no-copy-on-write
* synchronous-updates
See chattr and filesystem manual pages for details on those
attributes. More attributes for the spare part context can be
added on request. This Fixes#1233
In an OEM deployment that requested the creation of a swap
partition via <oem-swap> that swap partition was created
at first boot and was always the last partition on the disk.
This was required because it could not be placed before
any other partition without destroying those partition
contents. This process leaves the system in an inflexible
condition if the storage device can change its geometry
dynamicly as it's the case for SAN systems. The typical
deployment target for OEM images are SAN storage clusters
and it's cumbersome to resize the root partition if swap
is last.
This commit Fixes#1231 and changes the handling of swap if
requested via <oem-swap> as follows:
1. The swap space is created as part of the image build process
and no longer on first boot of the image via dracut code.
This increases the size of the non compressed .raw disk image
by the configured swap space size or the default. The
compressed versions are not affected since zero initialized
swap space compresses to almost no space. Deployment of
the image however also deploys the swap partition which
increases deployment time. For big swap configurations
it's advisable to switch off image verification via
oem-skip-verify. For very big swap configurations it's
also recommended to prevent kiwi from adding them as part
of the image and let them be created on first boot via
a systemd service that e.g places a swap file, or creates
a swap volume when possible such that the fexibility to
resize the rootfs is still available.
2. The setup of the swap space is now explicit. It's no longer
calculated by twice times RAM size because on newer machines
this could lead to huge numbers. Either the kiwi encoded
default swap size applies or the user configured value.
3. LVM based oem disks creates the swap space as logical volume.
The volume is created as part of the image build process
and no longer on first boot. The swap volume at build time
of the image is of a minimal size and gets resized on first
boot.
4. The move of the swap creation into the builder code also
handles swap per configured device persistency schema like
any other devices. This means by default swap is mounted
via by-uuid name and thus also Fixes#1259
The preparation to call zipl and the call itself were wrong.
For whatever reason the kernel image the initrd are moved
to another location prior to calling zipl. That move broke
the system because no kernel/initrd existed at the expected
place anymore. In addition the zipl call itself was issued
from a the wrong directory. Also no config file was written
as an after effect of the refactoring in Issue #1194. This
Fixes#1173 and bsc#1156694
