Remove the fallback clean up code, as it also may lead to images getting
removed right before we need to use them. Besides that, it should be
dead code by now since docuum should be running on all our jenkins nodes
to clean up old images based on last use date.
Don't delete images while they are being used, to fix these errors we
see from time to time in the middle of "docker build" on jenkins:
unknown parent image ID sha256:1b072e35048cd8b680eddabdc641ac678edb1184d222d5e7b3fbe0b3c333129a
This happens because "docker build" creates so-called dangling images
for each step processed of a Dockerfile. The "docker system prune" call
deletes these dangling images (among other things).
Remove the "docker system prune" call. We already have the docuum daemon
to deal with unused images (dangling and not dangling), it removes them
based on last use date so that the used space is always below a
configured limit. As it deletes images that haven't been used the
longest when it reaches the limit, it will not result in the problem
Besides images, "docker system prune" also removes unused containers
(instances of images created with 'docker run' without --rm) and
networks. Add "docker container prune" and "docker network prune"
commands to remove them from now on.
Also remove the redundant container removal logic (previous it was
redundant with "docker system prune", now redundant with "docker
After building packages with the OBS scripts, show their contents. This
allows easy checking if config files were correctly packaged etc. By
adding it here, it will also show up at the end of the related jenkins
jobs for gerrit verifications.
Let the gerrit-binpkgs and gerrit-pipeline-result jobs use the
"osmocom-gerrit" label. Having the -debian10/11 label previously was
useful to make sure this doesn't run on an ancient python version that
fails on f-strings. But since the debian9 builders were retired, this is
not needed anymore. These are the only jobs that used these, so now the
labels in the README can be simplified (following patch).
The docker container creates the trial dir and some files inside of it
as root, therefore the rm -rf trial-* line fails. Add logic to fix the
permissions before attempting to remove old trial-* and when the script
The virtual osmo-gsm-tester has been failing for a long time, but the
test results were still green. This was because the job failed before it
removed old test results, and so those were used. Remove the old results
as early as possible to prevent this in the future.
Build the docker image right before using it, instead of building it in
another jenkins job update-osmo-ci-on-slaves via
The logic in osmo-ci-docker-rebuild.sh was broken. I didn't realize at
the time that this image is only used for the virtual osmo-gsm-tester,
not the physical ones. But only the machines running the physical
osmo-gsm-tester have the /var/tmp/osmo-gsm-tester/state path. The
virtual osmo-gsm-tester isn't running on these machines but on generic
Building the image right before using it makes sense for this job, as it
is the only user of the image. If it was already built from the same
Dockerfile, a cached version is used.
Unable to find image 'osmocom-build/osmo-gsm-tester:latest' locally
Fixes: 9139e76b ("osmo-ci-docker-rebuild: don't always build osmo-gsm-tester")
Use literal style instead of folded style in the shell argument of
osmo-gsm-tester_virtual as otherwise new lines behave unintuitively.
Change it to "master" instead of "origin/master" to prepare for building
the docker image right before running it. The Makefile from
docker-playground.git picks up the OSMO_GSM_TESTER_BRANCH environment
variable and expects it to be in that format.
Remove the deb9 lxcs as they are being retired, and add two deb11 lxcs.
Coverity is quite big with 1.2 GiB, so deploying and upgrading this
takes some time. Having it on two hosts should be enough.
The tarball provided by coverity doesn't have the same version in the
name as in the directory inside the tarball.
Work around this by extracting it into a subdirectory and using glob to
match the directory name.
Coverity doesn't make the old version available for download and will
probably deprecate the previous version at some point. I'm about to
install it on new hosts, so upgrade to the latest version while at it.
Prepare to retire build2-deb9build-ansible. With changes in the script,
it doesn't need a persistent workspace anymore, it skips building source
tarballs if they exist on the remote server.
Depends: docker-playground I4b8c149c9cdbe7c613eea5d9be15794de5e5ddce
Remove the node attribute as these can run on any node. I've configured
the three deb11 nodes with "Usage: Use this node as much as possible"
instead of previously "Only build jobs with label expression matching
this node", so simply not specifying a node is now possible. Previously
it wouldn't have any node to run on as all of them were configured that
Don't attempt to build openbsc 1.4.0 and 1.4.1 in the
osmocom-build-old-tags-against-master job, as it fails to build with a
modern gcc with a potential null pointer dereference warning that gets
treated as error. This was fixed in 1.4.2 with openbsc patch
A more modern gcc gets used, as this job gets changed to not only run on
jenkins nodes with debian 9 installed, but also more modern versions.
This job does not run in docker, and since it still passes it doesn't
seem worth changing that now.
Use new ssh keys for ftp.osmocom.org with consistent naming scheme. It
was not possible to use @ in the ID name for new IDs, so I've used an
Allow overriding the ssh agent credentials per repository, and use
binaries_ftp.osmocom.org where needed. Use docs_ftp.osmocom.org instead
of 97634b9d-a867-4f6d-825d-1da99813e93d so it's clear what the key is
for. I generated a new one with the new ID.
Pass the ssh agent to docker for master builds, so we can publish
manuals and other files generated during the builds. This used to work
with the .ssh directory where each lxc had its own key. But we are
changing that to have the keys in jenkins and use the ssh agent, so it
works without additional configuration when adding new nodes/lxcs (such
as the two deb11 lxcs we just added).
Add the key to known hosts to prevent host key verification failed
errors when attempting to use the gerrit API from jenkins (to post
review comments) or when attempting to clone over ssh from gerrit (as
we do for some reason apparently).
With recent changes in open5gs, building the source packages we
[ 45s] ../lib/metrics/meson.build:52:4: ERROR: Automatic wrap-based subproject downloading is disabled
[ 45s] dh_auto_configure: error: cd obj-x86_64-linux-gnu && LC_ALL=C.UTF-8 meson .. --wrap-mode=nodownload --buildtype=plain --prefix=/usr --sysconfdir=/etc --localstatedir=/var --libdir=lib/x86_64-linux-gnu --libexecdir=lib/x86_64-linux-gnu returned exit code 1
Fix this by downloading all subprojects. This is now possible as I added
a meson.build file to the prometheus-client-c repository that open5gs
Send the mails to the mailing list rather than to only Harald and me, so
it's less likely that failures go unnoticed in case we are both
Make it consistent with all other jobs and don't set a display name.
This way the job will be sorted properly in the list of jobs and won't
weirdly jump to the top when loading the page.
Add a jenkins-job-builder configuration for the existing Osmocom_API job
and refactor it to not use the deprecated multi-scm plugin. Also update
the git clone urls.
Fix an error when the same Change-Id is used on multiple branches or in
theory projects. This is actually allowed by gerrit, and we use this
e.g. when backporting patches from master.
Use the project, change number (e.g. 30147) and patchset number (e.g. 2)
instead of the Change-Id.
+ ssh -p 29418 -l jenkins gerrit.osmocom.org gerrit review 4835a62cd88f0d69db76fb3bfd2df02176a91a6d --json
fatal: "4835a62cd88f0d69db76fb3bfd2df02176a91a6d" matches multiple patch sets