After the introduction of 'Add output for mis-configured element
scripts' we started seeing CI failures in tripleo where
instack-undercloud is being used (rocky/queens):
/usr/lib/python2.7/site-packages/diskimage_builder/lib/dib-run-parts: line 108: DIB_DEBUG_TRACE: unbound variable
INFO: 2019-12-02 16:24:33,423 -- ############### End stdout/stderr logging ###############
ERROR: 2019-12-02 16:24:33,423 -- Hook FAILED.
Let's make sure that by default the env variable is set
to 0.
Change-Id: I38c76c0edee436f1e7dd0c9a868cea1e6ee3271d
Closes-Bug: #1854904
When running under nodepool in a foreground, non-daemonized situation
without a tty (i.e. within a container) we're seeing this "wait" hang
indefinitely.
It is probably related to "outfilter.py" and output file descriptors,
although TBH we haven't completely root-caused it. I won't claim this
is a great solution, but it should hopefully let the dib process
finish and just die, where outfilter will disappear.
Change-Id: If78da54df3d4c240fee16aee4413ec554b37c1d6
I commonly get asked for help when people are attempting to create
local image elements and they cannot get them to work.
diskimage-builder silently ignores element scripts that it doesn't
find to it's liking, such as non-executable or files with extensions
(.sh is a common mistake).
This patch extends the '-x' tracing flag down to dib-run-parts and
will cause it to print out helpful messages when these files would
otherwise be silently ignored.
Examples:
Ignoring non-executable files: 10-do-not-run-me
Ignoring non-conforming filenames: 10-I-can-run.sh
I am not enabling these by default as they can create extra noise
and require additional filesystem IO to produce.
Change-Id: Ic804efca3015c199440b4b10da951d71a815c64f
This reverts commit a3e9e7f89e.
We still have some issues with vhd creation on RAX
In short, it appears that images fail to resize unless they have a
specific "creator" field. Revert this while we consider the options.
[1] https://bugs.launchpad.net/nova/+bug/862653
Change-Id: I2b6a3bfbfe28432fbb6a2ce4a0211939d224b8d5
The vhdutil utility is completely dead; the whole subsystem it relies
on was removed with [1] so it's not even vaguely possible to keep it
up-to-date.
I took the .raw images on a nb and used the qemu-img there (so Xenial)
and generated some VPC images; uploaded them to rackspace and the all
seemed to boot fine. If there was a problem, maybe it's been fixed on
either the qemu or RAX side in the previous few years.
Thus swith to qemu-img to generate the vhd images too.
[1] https://xenbits.xen.org/gitweb/?p=xen.git;a=commit;h=5c883cf036cf5ab8b1b79390549e2475f7a568dd
Change-Id: I3099d2ebb958370fcec623087a093b2c8dbdc6c4
Add a new environment variable $DIB_GZIP_BIN allowing builders to
specify a different gzip (such as pigz) to be used when compressing
tgz images.
Change-Id: Ifb617568140a149e2fda241e07ff8a59429e6697
As noted in the change, 7fd52ba841
increased the size of the EFI partition considerably. This has meant
that our padding upwards of the disk size is insufficient and EFI
builds (arm64 in particular) is failing due to out-of-disk errors
during final image operations like installing kernels.
Similar to the discussion we had in
I65fa13a088eecdfe61636678578577ea2cfb3c0c, this feels a bit ugly
because we're mixing logic here with sizes specified in block-device
config files. But it boils down to the same problem; we are
calculating the disk size here and passing it to the block-layer, so
unless we want to make large changes to the status quo about where
these sizes are calculated, small adjustments here are the most KISS
solution.
Thus we check if we have selected the EFI bootloader element, and thus
assume there will be a large system EFI partition and expand the disk
size accordingly.
Change-Id: Ifa05366c2f2b95259f3312e4dde8c85347075ba1
python 3.6 warns about regexes like:
DeprecationWarning: invalid escape sequence \+
I noticed that debugging a trove job and it really led me in the wrong
way. Fix this with making it a raw string.
Change-Id: I58ee1a49d62316c6c3f0588832c97f659f7e460b
This patch removes the check and default for rhel 8 requiring
xfs filesystem as rhel 8 images can successfully be built with
ext4 filesystems.
Change-Id: I1a6bfa26324fd43ae0c77c2c977dda0dd56e26e5
Make a version-less RHEL element to handle both '7' and '8' DIB_RELEASE.
The element usage should align with other elements which operate in the
same way such as the Fedora element.
Additionally, this patch adds support for RHEL8 that operates with
Python 3.
As of now, users of diskimage-builder will still be able to use the
'rhel7' element, or migrate to 'rhel' and specify their respective
DIB_RELEASE value.
* mount the xfs file-system for extraction as read-only. vaguely
based on explaination in [1] and the fact we only read the image
data into a tar, so can ignore this.
XFS (dm-1): Superblock has unknown read-only compatible features (0x4) enabled.
* Use the redhat system python as the dib-python version. dib was
ahead of it's time making an abstracted python interpreter for
system work ;) the system python should work for running the various
dib element scripts.
[1] https://unix.stackexchange.com/questions/247550/unmountable-xfs-filesystem
Redhat-Bug: https://bugzilla.redhat.com/show_bug.cgi?id=1700253
Co-Authored-By: Ian Wienand <iwienand@redhat.com>
Change-Id: I90540675c70bb475d9db2ae24f81c648a31f3f95
I want to use the new --image-extra-size flag[1] but my use-case
calls for megabyte granularity of this value. Rather than adding
60% to an 800MB image, maybe I only want to add 100 or 200MB, etc.
[1] https://review.opendev.org/#/c/655127/
Change-Id: I8fb9685d60ebb1260d5efcf03c5c23c561c24384
Currently diskimage-builder supports two ways to specify the image
size. One is defining a fixed image size using DIB_IMAGE_SIZE, the
other one is auto-detection while adding a security margin of 60% as
free space. This means when building larger images (e.g. >100GB) with
unknown size upfront we end up with much wasted space, IO and network
traffic when uploading the images to several cloud providers. This can
be optimized by adding a third way by defining DIB_IMAGE_EXTRA_SIZE to
specify the free space in GB. This makes it possible to easily build
images of varying sizes while still minimizing the overhead by keeping
the free space constant to e.g. 1GB.
Change-Id: I114c739d11d0cfe3b8d8abc6df5ff989edfb67f2
In many cases, the statically sized 64MB journal is far below the
e2fstools default calculation[0] which calls for a 64MB journal only
on filesystems smaller than 16GB. On bare metal in particular, the
correct default journal size will often be in the 512MB-1GB range.
Since we cannot know what the target system is, this should be a
tunable parameter that the user can set depending on the intended
image usage.
Add a DIB_JOURNAL_SIZE envvar and --mkfs-journal-size parameter
to the image creation so users can override the default journal
size.
[0] https://git.kernel.org/pub/scm/fs/ext2/e2fsprogs.git/tree/lib/ext2fs/mkjournal.c#n333
Change-Id: I65fa13a088eecdfe61636678578577ea2cfb3c0c
This is only one line, but it takes a lot to untangle ... basically
the current "correct" path is:
---
mk_build_dir()
-> sets trap trap_cleanup EXIT
... stuff ..
mount_proc_dev_sys
-> mounts $TMP_MOUNT_PATH/<proc,dev.sysfs>
pre-finalise.d
finalise.d
unmount_image $TMP_BUILD_DIR/mnt # nb == $TMP_MOUNT_PATH
-> unmount_dir()
-> recursive unmount everything inside TMP_MOUNT_PATH
TMP_IMAGE_PATH=$(dib-block-device getval image-path)
export TMP_IMAGE_PATH
dib-block-device umount
dib-block-device cleanup
... actually cleanup directories ...
---
Our current failure exit trap does:
---
dib-block-device umount
unmount_image
...
---
Note this is the *opposite* of what is done in the correct exit path.
In the failure case, if a script fails in the finalise stages it leads
to /proc, /sys, /dev etc. still being mounted inside the image; the
"dib-block-device umount" call doesn't know anything about these
mounts and tries to unmount the parent directory, and we get a hard
failure with a busy mount, and all the mounts are subsequently leaked.
Note that "unmount_dir", which is ultimately called by
"unmount_image", already knows to skip those mounts that
"dib-block-device umount" manages (this is the DIB_MOUNTPOINTS list).
This is further evidence it should be called *before* the
dib-block-device umount.
Change-Id: Ibef3ce9d1167b9c4ff3d5717b113cd3ed374f5e3
The path $TMP_BUILD_DIR/mnt becomes the / inside the chroot during
the chroot phases of diskimage-builder. Previously this path was being
created using the account running diskimage-builder. This account may
not be valid inside the chroot. This causes path validation, when running
on a Ubuntu bionic host, to fail.
This patch chown's the $TMP_BUILD_DIR/mnt to root.root to make sure
that / is owned by a valid account inside the chroot.
Change-Id: Ifedc136baa67c7952942aed2c8cb1041902fef91
Closes-Bug: 1811113
It looks like we dropped running these probably when we moved the
elements around. For testtools to find the test scripts we need to
add the __init__.py files to make the directories look like modules.
Also prevent copying any .pyc or cache files in as hooks.
Change-Id: I66d5f6ee62cc4d9ee14c64e819b4db57d035d09f
I'm not really sure why I originally had --logfile also log to stdout
in I202e1cb200bde17f6d7770cf1e2710bbf4cca64c, but it seem
counter-intuitive (indeed, I just tripped myself up thinking that in a
devstack job "--logfile" would put the logs into a separate file and
avoid the stdout logging, and I wrote it!).
Make it so specifying a --logfile puts dib into quiet mode for stdout.
Explicitly overriding DIB_QUIET will allow both if someone wants that.
Change-Id: I3279c9253eee1c9db69c958b87a0ce73efc0be9b
While trying to get docker image pre-caching to work we couldn't get a
docker daeomon to run within the chrooted environment. However we got
docker running with the help of bwrap outside of the chrooted
environment. The only option so far for this is the block-device.d
phase. But this has the problem that it runs after the image size has
been calculated. This leads to broken builds if the docker images
being pulled are big.
This can be solved by adding a post-root.d phase that runs outside the
chroot but before the image size calculation.
Change-Id: I36c2a81e2d9f5069f18ce5b0d52c5f1c7212c3ae
This is a lot of very low value noise in the logs as these iterate
through all the elements (often doing nothing). Turn it down and add
an echo so we just see what elements it is working on.
Change-Id: I0687de4722766189db9d4a7bd7d3cfb45d387b62
In exploring Gentoo caching, it was realised that we have no way to
bind mount the cache into the finalised image for the finalise.d
phases.
By adding a pre-finalise.d phase that runs outside the chroot, we can
mount outside things into the hierarchy at $TMP_BUILD_DIR/mnt which
are then seen by the in-chroot finalise.d phase.
This is similar to the pre-install phase
Change-Id: I9d782994843383ddf90f62c40498af9925fd9558
Some minor things after looking at these parts.
The dib-run-parts element doesn't do any of the copying any more, so
these comments are wrong.
The reason for the multiple mounts in the bind mount was non-obvious
to modern eyes (as util-linux has handled this for some time).
Formatting fix for the rst
Change-Id: Idb4c9ff32c49aced2c68a5c905bf7a8b2832a5a2
Redirecting our output through outfilter.py is inherently a bit racy,
since the disk-image-create process will exit, and then you might get
outfilter.py flushing any remaining output as it closes.
On an interactive prompt this might lead to final output overwriting
the prompt, etc. This can be a bit confusing when you start running
things in a loop.
If we save the original fd, then on the exit path close the redirected
fd's and wait a little bit for final output (as a result of the
close), we get a more consistent output.
Change-Id: I8efe57ab421c1941e99bdecab62c6e21a87e4584
Strip everything before "site-packages" in the output filename for the
PS4 prompt. This makes the line in debug logs significantly shorter
as we don't have the full virtualenv path every single time. The
important thing -- the file being called in the lib/ dir, is retained.
Change-Id: I00706b6f6c0425c7795f997c08ceda3374dc84b5
When switching to using log-file capture, we're getting
[gentoo/build-succeeds] outfile.write(ts_line.encode('utf-8'))
[gentoo/build-succeeds] UnicodeEncodeError: 'utf-8' codec can't
encode character 'udcc5' in position 59: surrogates not allowed
Use surrogateescape [1] on the output to avoid this
[1] https://www.python.org/dev/peps/pep-0383/
Change-Id: I2c2c537296edfa5a8fe661a41bd5bfb3bfcf57e3
The behavior of test -e and [[ -e against broken symlinks is to fail
even if the symlink exists. However we want to test if the link exists
or if there is a file in that location. Therefore switch from test -e to
test -L and test -f to check if the file or link exists regardless of
link target validity.
Change-Id: I84a9b6731eccf950707be50aef464a2de1e33e8e
systemd-resolved has a new behaviour in bionic, in that if there is no
/etc/resolv.conf file when it installs, it assumes it is a fresh
system and makes /etc/resolf.conf a symlink into its compatability
files.
dib ends up saving & restoring whatever /etc/resolv.conf we have after
the inital chroot creation, which may not be what we want -- in the
above case it restores the system-resolved symlink. For
openstack-infra, we use unbound and want simply "127.0.0.1" in a
/etc/resolv.conf file [1].
Formalise the ability to save specific contents into the final image.
Add documentation, and a note in the code that it's an external
interface.
I would have preferred to namespace the .ORIG file with DIB_ or
similar, but this unofficial interface has already escaped into the
wild. Leave it as is for simplicity.
[1] Note that systemd-resolved will obey /etc/resolv.conf as you would
expect, if file exists.
Change-Id: Ie0e97d8072e2b21a54b053fa6fb07b62960c686d
We exit in several places and don't restore tracing. Previously in
nodepool we relied on the default fallback, which did restore the
tracing. Since we now use the MBR config file, we take the different
exit path without it and the debugging output is incomplete.
Change-Id: I586fc95517926025705ce376ec5c4aaf4122773f
This moves the block-device default out of the "vm" element and into a
selection of other elements. There's "mbr" which retains the status
quo. There's an EFI version that has the boot/grub partitions as
required. In between there's the GPT only version, which is useful
for architectures like power without EFI, but still want possible
larger disks using GPT.
Change-Id: I4a566a97d073fc0dda0ab2494ac988fe015800a9
The current implementation - as introduced in
Iee44703297a15b14c715f4bfb7bae67f613aceee - has some shortcomings / bugs,
like:
* the 'grep' check is too sloppy
* when /dev/pts is already mounted multiple times the current implementation
fails:
$ mount | grep devpts | sed 's/.*(\(.*\))/\1/'
rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000
rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000
rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000
* code duplication
* Using the undocumented and non-robust output
of 'mount'.
This patch fixed the above problems.
Change-Id: Ib0c7358772480c56d405659a6a32afd60c311686
Signed-off-by: Andreas Florath <andreas@florath.net>
This is a continuation for f2cc647dae ("diskimage_builder: lib:
common-functions: Fix options for devpts mount"). We also need to
respect the devpts mount options when the dib elements are mounting
this virtual filesystems themselves.
Change-Id: Iee44703297a15b14c715f4bfb7bae67f613aceee
Commit cebfcf85f9 ("Use -t devpts for
/dev/pts mounts") switched from using '--bind' to '-t devpts' for
mounting the /dev/pts virtual filesystem. However, mounting devpts to
another location also affects the host's /dev/pts mountpoint. Since we
are now mounting devpts without options we end up with the following one
on openSUSE
devpts on /dev/pts type devpts (rw,nosuid,noexec,relatime,mode=600,ptmxmode=000)
instead of the one we want
devpts on /dev/pts type devpts (rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000)
The missing gid=5 options results to boot problems for virtual machines
So in order to fix that, we need to use the existing devpts options for
/dev/pts so we don't lose them in the new mount.
Change-Id: I17f2c2bb96b807f8dbc07185ae0147bff3230f92
In a couple of places we use flock for critical sections, but we leave
lockfiles around in various locations which can be confusing.
Introduce DIB_LOCKFILES global (under ~/.cache/dib/lockfiles) and
write lockfiles in there.
Fix up removal of the lockfile in the yum path; we just want to make
sure we cleanup the .rpmmacros file, but we don't need to remove the
lockfile as well.
Co-Authored-By: Andreas Florath <andreas@florath.net>
Change-Id: Ie810b2836be521325afe923708d046112e1e1e20
Currently a bind is used when mounting /dev/pts in chroot.
This leads to problems - especially when running DIB in parallel:
It was observed that the /dev/pts mount vanishes from the host
system.
This patch uses '-t devpts' - as it is done for /sys and /proc -
for handling /dev/pts.
Change-Id: Id7775ae6fca6502af800e7b73a00862ef320206b
Signed-off-by: Andreas Florath <andreas@florath.net>
The call to fstrim in disk-image-create is currently useless, because
at the time this is called, the file systems were already umounted by
the block device layer.
The current implementation of the block-device mount plugin does not
call fstrim at all: resulting in larger image sizes.
This patch removes the useless fstrim call from the disk-image-create
script and moves this into the block-device mount.py.
The resulting image might be much smaller. Example: Ubuntu Xenial
with some elements; once with and once without this patch:
-rw-r--r-- 1 dib dib 475661824 Sep 16 06:43 ubuntu-xenial-without-fstrim.qcow2
-rw-r--r-- 1 dib dib 364249088 Sep 16 09:30 ubuntu-xenial-with-fstrim.qcow2
Change-Id: I4e21ae50c5e6e26dc9f50f004ed6413132c81047
Signed-off-by: Andreas Florath <andreas@florath.net>
Under certain environments, this timeout was causing failures
because it was too short. Increasing to 10, to give time to
perform the specified tasks.
Change-Id: I01dd3553f38e1137b2fcb04b4ee12202be3ad1a8
We replace the base resolv.conf with an "outside" copy so that
resolving works when we're in the chroot.
Installing resolvconf package modifies the in-chroot resolv.conf to a
symlink (to /var/run) which it wants maintained in the final image.
We have the existing "immutable" check for a created resolv.conf file,
but no eqivalent for a symlink.
This adds a check to see if the resolv.conf is a symlink and leave it
alone if it is, assuming it has been re-created in the chroot.
I have tested this with ubuntu-minimal+resolvconf with
dhcp-all-interfaces and the system seems to work with resolvconf
working correctly.
Change-Id: Idd5a26e9d55979bd951577d5b098ed4bfba91ad3
