lp:~dbungert/curtin

Author: Dan Bungert
Author Date: 2025-09-18 21:26:32 UTC

curthooks: install zfs-dracut as needed

When using zfs, installing zfs-initramfs is only correct if the target
is using initramfs-tools. Handle the dracut case.

LP: #2125164

Author: Dan Bungert
Author Date: 2025-08-26 00:21:59 UTC

.

Author: Dan Bungert
Author Date: 2025-08-25 23:06:11 UTC

concept

Author: Dan Bungert
Author Date: 2025-08-22 20:12:15 UTC

distro: ensure_one_kernel - fix metapackage + no kernel install

In the case where we list metapackages in the before list for
ensure_one_kernel(), and we haven't actually installed a kernel during
the yield, there is nothing to remove. Strict set equality check
doesn't handle that correctly.

Author: Olivier Gayot
Author Date: 2025-04-15 12:35:08 UTC

zfs: increase size of keystore

On plucky, a 20MiB keystore does not allow the install to complete. For some
reason, cryptsetup seems to place the data segment at offset 32MiB instead of
16MiB (which was the case previously).

This results in warnings & errors:

When running crtypsetup luksFormat:

> Device /dev/zvol/rpool/keystore is too small for activation, there is no remaining space for data.

When runnign cryptsetup open:

> Requested offset is beyond real size of device /dev/zvol/rpool/keystore

LP: #2107381

Signed-off-by: Olivier Gayot <olivier.gayot@canonical.com>
(cherry picked from commit 148b02f5f4027ca508727110604e376f325a1078)

Author: Olivier Gayot
Author Date: 2025-04-15 12:35:08 UTC

zfs: increase size of keystore

On plucky, a 20MiB keystore does not allow the install to complete. For some
reason, cryptsetup seems to place the data segment at offset 32MiB instead of
16MiB (which was the case previously).

This results in warnings & errors:

When running crtypsetup luksFormat:

> Device /dev/zvol/rpool/keystore is too small for activation, there is no remaining space for data.

When runnign cryptsetup open:

> Requested offset is beyond real size of device /dev/zvol/rpool/keystore

LP: #2107381

Signed-off-by: Olivier Gayot <olivier.gayot@canonical.com>
(cherry picked from commit 148b02f5f4027ca508727110604e376f325a1078)

Author: Dan Bungert
Author Date: 2025-06-03 20:27:19 UTC

Revert "curthooks: remove all explicit calls to update_initramfs"

This reverts commit 1a968edbe0caaf9113542ed38747a2ed83d8bf27.

Author: Dan Bungert
Author Date: 2025-06-03 16:39:29 UTC

util: link to LP bug for flock()

Author: Dan Bungert
Author Date: 2025-03-13 23:32:50 UTC

unittests: remove usage of imp module

`imp` has been removed as of python3.12.
https://docs.python.org/3.12/library/imp.html

We could migrate to importlib, but this is about python2 compat and we
don't need that anymore, so just delete stuff.

(cherry picked from commit 1667812bc4b2393776a8ee32d558b6ffe37d3b11)

Author: Dan Bungert
Author Date: 2025-05-27 17:44:37 UTC

vmtest-system-setup +zfsutils-linux

Author: Dan Bungert
Author Date: 2023-05-16 17:59:37 UTC

integration: add run_bm(..., debug=True) helper

The debug argument for run_bm allows for running curtin outside of the
integration test itself, but with all the setup being done for you by
the integration test. This means that custom tweaks can be done to the
curtin code before running, setting breakpoints, etc.

Author: Dan Bungert
Author Date: 2025-04-09 20:51:59 UTC

curthooks: set FK env vars for dpkg-reconfigure of the kernel

LP: #2106682

Author: Dan Bungert
Author Date: 2025-03-31 21:03:34 UTC

several: fix flake re unassigned globals

Author: Dan Bungert
Author Date: 2025-03-26 21:40:07 UTC

curthooks: move kdump enablement after initramfs generation

Author: Dan Bungert
Author Date: 2025-03-24 22:33:30 UTC

control: build-dep on python3-packaging

Author: Dan Bungert
Author Date: 2025-03-24 17:54:50 UTC

whitespace increment to trigger ci

Author: Dan Bungert
Author Date: 2025-03-21 16:55:07 UTC

curthooks: rename kernel event

initramfs config sounds more specific than what this really does. Now,
this runs kernel postinstall hooks for Debian-like systems, providing
the updated initrd functionality along with other expected behaviors.

Author: Dan Bungert
Author Date: 2025-03-14 00:04:33 UTC

curthooks: reconfigure kernel after initrd generated

Kernel postinstall scripts do several helpful things that we don't want
to duplicate code for here in curthooks. Run those scripts.

Done with LP: #2101831 in mind but expected to be more broadly useful.

Author: Dan Bungert
Author Date: 2025-03-14 00:04:33 UTC

curthooks: reconfigure kernel after initrd generated

Kernel postinstall scripts do several helpful things that we don't want
to duplicate code for here in curthooks. Run those scripts.

Done with LP: #2101831 in mind but expected to be more broadly useful.

Author: Dan Bungert
Author Date: 2025-03-13 20:21:38 UTC

curthooks: move zipl invocation after update-initramfs

To match the behavior of the zz-zipl script in s390-tools, zipl should
be invoked after the initrd is in place.

Author: Dan Bungert
Author Date: 2025-03-10 19:11:48 UTC

Revert "Add doc/_build to the .gitignore list"

This reverts commit 8bde61c7531301bd91f7fdbbf162faac47b0c722.

Author: Dan Bungert
Author Date: 2025-01-10 23:15:33 UTC

tox: switch to pytest

(cherry picked from commit f9e77d3eef78d4ada60a1da1dda67fce363fc8e2)

Author: Dan Bungert
Author Date: 2025-01-10 23:15:33 UTC

tox: switch to pytest

(cherry picked from commit f9e77d3eef78d4ada60a1da1dda67fce363fc8e2)

Author: Simon Glass
Author Date: 2025-01-07 17:19:42 UTC

Add initial support for extlinux bootloader

Add a basic implementation of this, mirroring the u-boot-update script.
For now, this lacks several features which require schema additions.

The main question here is how to find the correct root= argument for
Linux.

Signed-off-by: Simon Glass <simon.glass@canonical.com>

Author: Dan Bungert
Author Date: 2025-01-27 16:14:59 UTC

distro: fix purge of packages from a PPA

As suggested by LP user icecream95, the sequence purge - autoremove may
fail if there are PPA packages involved.

LP: #2091689

Author: Dan Bungert
Author Date: 2025-01-10 23:15:53 UTC

tox: sitepackages = false

If we pull python3-apt from git we can drop sitepackages, which helps
insulate against CI problems. We need libapt-pkg-dev for that though.

Author: Dan Bungert
Author Date: 2025-01-06 18:08:30 UTC

README: add unnecessary whitespace

Author: Dan Bungert
Author Date: 2024-12-05 17:12:54 UTC

curthooks: fix for py3.5

Fix a syntax issue, get unit tests running again. Pulling test
requirements is pretty broken trying to run on Xenial so the test
strategy is archive packages and
PYTHONPATH=. py.test-3 tests/unittests/

Author: Dan Bungert
Author Date: 2024-10-30 09:35:43 UTC

readthedocs: add basic template

This is a required file for modern RTD. Try the basic template to
start.

Author: Dan Bungert
Author Date: 2024-10-27 11:29:31 UTC

curthooks: fix early exit on non-grub cases

Author: Chris Peterson
Author Date: 2024-09-14 00:10:53 UTC

Introduce the kernel-crash-dumps builtin hook

Introduce a new hook for builtin_curthooks to configure kernel crash
dumps on the target system. The configuration key is
`kernel-crash-dumps` and adds the following new section to the
curthooks configuration:

    kernel-crash-dumps:
        enabled: bool | None

By default, `enabled` is `None` will cause kernel crash dumps
to be dynamically enabled on the target system if the enablement
script provided by kdump-tools is found in:

/usr/share/kdump-tools/kdump_set_default

The enablement script will inspect the system and either enable or
disable kernel crash dumps. Users can also specify `True` or `False`
to unconditionally enable or disable kernel crash dumps, respectively.

(cherry picked from commit 40cae5c60fa9f4c495c7f61cde28862175f93ce2)

Author: Dan Bungert
Author Date: 2024-09-18 21:30:02 UTC

config: use older typing syntax for py3.8

Author: Dan Bungert
Author Date: 2024-07-24 20:45:12 UTC

distro: fix handling of pre-installed OEM kernel

In the case of a pre-installed OEM kernel, we need to handle removals
differently. Since Subiquity is managing OEM kernel installs, we need
that outside input to to know what the "before" state of the kernels is,
and the existing handling of ensure_one_kernel works well with that if
we just pass along that "before" state.

Have Subiquity feed the initial state (calculated before attempting to
install OEM kernels) will handle the OEM case - we start and finish with
one kernel so leave it alone - along with the two kernel case - we have
some other kernel pre-installed then Subiquity installs the OEM kernel,
and the remove list is that pre-installed kernel.

Author: Dan Bungert
Author Date: 2024-07-19 21:11:37 UTC

curthooks: support explicit kernel removal

Author: Dan Bungert
Author Date: 2024-07-22 19:38:26 UTC

maybe

Author: Dan Bungert
Author Date: 2024-07-17 20:26:15 UTC

curthooks: attempt to skip kernel installs

For the OEM case, it helps if we don't redundantly try to install
kernels already installed, as we may not have that repo in our list
right this second.

Author: Dan Bungert
Author Date: 2024-07-12 19:10:36 UTC

integration: quick_zero test

LP: #2061073

Author: Dan Bungert
Author Date: 2024-07-11 19:31:10 UTC

vmtests: fix basic_scsi.yaml running out of space

Author: Dan Bungert
Author Date: 2024-07-11 19:31:10 UTC

vmtests: fix basic_scsi.yaml running out of space

Author: Dan Bungert
Author Date: 2024-07-09 21:42:44 UTC

tox: unpin flake8 version, sort

Author: Dan Bungert
Author Date: 2024-07-01 21:14:59 UTC

README: note about github

Author: Dan Bungert
Author Date: 2024-06-06 23:57:28 UTC

distro: ensure_one_kernel debug logs

Author: Dan Bungert
Author Date: 2024-05-29 21:03:06 UTC

curthooks: support kernel pkg removal

Author: Dan Bungert
Author Date: 2024-05-22 20:44:41 UTC

curthooks: create and refactor to KernelConfig

Author: Dan Bungert
Author Date: 2024-04-15 18:34:20 UTC

default GRUB_CMDLINE_LINUX_DEFAULT to include quiet/splash

Author: Dan Bungert
Author Date: 2024-04-13 17:25:54 UTC

util: log release of flock

Author: Dan Bungert
Author Date: 2024-03-20 03:40:26 UTC

block/zfs: flock for ZFS

Add flock calls around encrypted ZFS creation areas

LP: #2057661

Author: Dan Bungert
Author Date: 2024-03-20 19:05:31 UTC

lint: fix whitespace

Author: Dan Bungert
Author Date: 2024-03-08 18:55:06 UTC

Revert "Bind ischroot->true in ChrootableTarget and --mount-proc in unshare"

This reverts commit c5bb06b9645dfa7c69b7f54f860c68c9d8f560d7.

Author: Dan Bungert
Author Date: 2024-02-29 23:47:25 UTC

do we need this for zpool export?

Author: Dan Bungert
Author Date: 2024-02-28 23:15:32 UTC

util: human2bytes actually return int

Author: Dan Bungert
Author Date: 2024-02-23 00:24:46 UTC

dm_crypt: tests related to keyfile removal

Author: Dan Bungert
Author Date: 2024-02-21 19:14:55 UTC

zpool: create keystore encrypted setups

Add support for the 'luks_keystore' style encrypted ZFS install. Native
ZFS encryption is supported by a small dm_crypt dataset containing the
real key. This is a functional transcription of the way that Ubiquity
has implemented encrypted guided ZFS.

(cherry picked from commit 5bf15d6373def082104d64a352c53ed079b537dd)

Author: Dan Bungert
Author Date: 2024-02-20 22:14:30 UTC

block_meta: improve get_volume_uuid

blkid can return 3 different types of UUIDs, and they aren't
interchangable. Clarify which one we're using.

Also, leverage the existing blkid() wrapper to do so.

Author: Dan Bungert
Author Date: 2024-02-20 22:58:09 UTC

better

Author: Dan Bungert
Author Date: 2023-05-16 17:59:37 UTC

integration: add run_bm(..., debug=True) helper

The debug argument for run_bm allows for running curtin outside of the
integration test itself, but with all the setup being done for you by
the integration test. This means that custom tweaks can be done to the
curtin code before running, setting breakpoints, etc.

Author: Dan Bungert
Author Date: 2023-06-08 18:09:53 UTC

maybe

Author: Dan Bungert
Author Date: 2024-02-08 16:05:53 UTC

block: fix kname kernel ref

Author: Dan Bungert
Author Date: 2024-02-06 23:52:44 UTC

block_meta: fix fstab path to cryptoswap

Curtin has this system by which it looks for the most robust path to the
device. Sadly, the existing mechanism fails for cryptoswap.

At install time, this device shows up with an ID that looks like
CRYPT-LUKS2 and insted ends up with CRYPT-PLAIN but the underlying
device remains luks (has the header).

This change appears to not affect other cases like LVM on LUKS, based on
regression testing of the existing Subiquity LVM+LUKS guided install
type.

I believe the name appropriate as a selection mechanism because we
already know it needs to be unique.

Author: Dan Bungert
Author Date: 2024-02-06 23:52:44 UTC

block_meta: fix fstab path to cryptoswap

Curtin has this system by which it looks for the most robust path to the
device. Sadly, the existing mechanism fails for cryptoswap.

At install time, this device shows up with an ID that looks like
CRYPT-LUKS2 and insted ends up with CRYPT-PLAIN but the underlying
device remains luks (has the header).

This change appears to not affect other cases like LVM on LUKS, based on
regression testing of the existing Subiquity LVM+LUKS guided install
type.

I believe the name appropriate as a selection mechanism because we
already know it needs to be unique.

Author: Dan Bungert
Author Date: 2023-11-11 13:32:40 UTC

block_meta: dm_crypt extensions for cryptswap

Author: Dan Bungert
Author Date: 2023-11-12 00:09:53 UTC

notes

Author: Dan Bungert
Author Date: 2023-10-05 23:51:04 UTC

install: do not log merged config

Author: Dan Bungert
Author Date: 2023-09-26 21:01:34 UTC

Revert "vmtests: add mantic after jammy"

This reverts commit abebab1e1e30a16d4be891a051060e81254ff0c5.

Author: Dan Bungert
Author Date: 2023-09-20 20:57:46 UTC

thoughts on devsync

Author: Dan Bungert
Author Date: 2023-09-19 19:01:20 UTC

vmtests: make loading error verbose

Author: Dan Bungert
Author Date: 2023-09-13 23:08:44 UTC

extract: log source information

If you try hard enough and follow mount commands backwards you can
surely get this same information, but when analyzing what the extract is
actually doing it's nice to just be able to see what the image stack is.

Author: Dan Bungert
Author Date: 2023-09-13 01:12:32 UTC

storage_config: handle partitions on 4k disk

In LP: #2035034, a user reported that the layout of existing
partitions was showing as 1/8th the correct size. Update storage_config
to no longer hard-code a 512 byte sector and read the value instead.

Author: Dan Bungert
Author Date: 2023-08-30 22:37:35 UTC

swap: add logging

The suggested swap size code just sort of returns 0 sometimes and
figuring out why is a complicated analysis of several variables.

Provide diagnostic info on those values, and the reason for the
decision.

Author: Dan Bungert
Author Date: 2023-08-30 22:37:35 UTC

swap: add logging

The suggested swap size code just sort of returns 0 sometimes and
figuring out why is a complicated analysis of several variables.

Provide diagnostic info on those values, and the reason for the
decision.

Author: Dan Bungert
Author Date: 2023-08-25 18:53:16 UTC

docs: link to sources.list manpage for apt deb822

Author: Dan Bungert
Author Date: 2023-08-24 17:41:02 UTC

zpool: allow selective default property override

zpool fs or pool properties may now use underlying fs defaults with
key: null.

Author: Dan Bungert
Author Date: 2023-08-23 20:45:50 UTC

zfs: only use default props if none are supplied

Author: Dan Bungert
Author Date: 2023-08-03 01:39:12 UTC

several: drop python3-mock

Author: Dan Bungert
Author Date: 2023-08-03 00:10:39 UTC

vmtest-system-setup: update for latest mantic

-pep8, not available
+parameterized, in use by tests

Author: Dan Bungert
Author Date: 2023-07-25 16:33:24 UTC

swap: support avail parameter

Add an avail parameter to allow swap size projections based on expected
available space. Add unit tests for existing and new (avail) usage.
Add docstring to `suggested_swapsize()`.

Author: Dan Bungert
Author Date: 2023-07-25 16:33:24 UTC

swap: support avail parameter

Add an avail parameter to allow swap size projections based on expected
available space. Add unit tests for existing and new (avail) usage.
Add docstring to `suggested_swapsize()`.

Author: Dan Bungert
Author Date: 2023-06-05 21:42:38 UTC

env hacks

Author: Dan Bungert
Author Date: 2023-05-25 20:23:30 UTC

several: fix lints

Author: Alexsander de Souza
Author Date: 2023-05-25 19:48:55 UTC

add python3-attr as dependency

(cherry picked from commit 1019040c497bdbf96613d7eb0dc682f92275d5b0)

Author: Alexsander de Souza
Author Date: 2023-05-25 19:48:55 UTC

add python3-attr as dependency

(cherry picked from commit 1019040c497bdbf96613d7eb0dc682f92275d5b0)

Author: Dan Bungert
Author Date: 2023-05-23 21:01:34 UTC

block: move ENXIO workaround up a level

Author: Dan Bungert
Author Date: 2023-05-19 18:53:00 UTC

block: only use no-tell-kernel if sfdisk supports

Author: Dan Bungert
Author Date: 2023-05-17 00:23:39 UTC

block: retry loop post dd image

On Bionic, I observed that the volume group present in the test image
would often not show up. This retry loop helps.

Author: Dan Bungert
Author Date: 2023-05-17 18:51:50 UTC

integration: fix tests with empty ptable

Older sfdisk seems to not like a completely empty ptable while setting
these attributes. Adding an unused partition is enough to allow
focusing on the task at hand.

Author: Dan Bungert
Author Date: 2023-05-17 18:51:50 UTC

integration: fix tests with empty ptable

Older sfdisk seems to not like a completely empty ptable while setting
these attributes. Adding an unused partition is enough to allow
focusing on the task at hand.

Author: Dan Bungert
Author Date: 2023-05-15 23:52:36 UTC

requirements: add missing content

Author: Dan Bungert
Author Date: 2023-05-19 18:53:00 UTC

block: only use no-tell-kernel if sfdisk supports

Author: Dan Bungert
Author Date: 2023-05-17 17:53:37 UTC

unittest: fix tests that scan log results

Author: Dan Bungert
Author Date: 2023-05-19 21:17:33 UTC

several: support older attrs version

Author: Dan Bungert
Author Date: 2023-04-26 11:23:31 UTC

more deps

Author: Dan Bungert
Author Date: 2023-04-18 14:31:48 UTC

block: ntfsresize with force options

To match Ubiquity behavior, we should run ntfsresize with the force
options and sending 'y\n' on input.

(cherry picked from commit 6caa799551b19dc85cb1e901e013b12be326aff1)

Author: Dan Bungert
Author Date: 2023-04-18 14:31:48 UTC

block: ntfsresize with force options

To match Ubiquity behavior, we should run ntfsresize with the force
options and sending 'y\n' on input.

Author: Dan Bungert
Author Date: 2023-02-07 17:56:37 UTC

.

Author: Dan Bungert
Author Date: 2023-02-07 01:40:09 UTC

integration: compare against sysfs results

Ignoring size on extended partitions, sysfs seems to match what sfdisk
is providing.

Author: Dan Bungert
Author Date: 2023-02-06 20:06:04 UTC

swap: clarify debug message

This log message is shown, among other scenarios, if a swap partition is
used, which is very confusing.
Clarify that we're talking about swap file.

Author: Dan Bungert
Author Date: 2023-01-28 00:36:24 UTC

dm_crypt preserve open

Author: Dan Bungert
Author Date: 2023-01-24 18:24:26 UTC

vmtests: drop python3-attr dep

Author: Dan Bungert
Author Date: 2023-01-24 17:07:35 UTC

block: remove fstrings

Author: Dan Bungert
Author Date: 2023-01-18 18:10:25 UTC

test