Feature #4563
closedinstall test for centos packages (like the debian install test)
100%
Description
Create jenkins jobs, that install all centos packages from the official repositories for nightly and latest.
Related issues
Updated by osmith almost 4 years ago
- Related to Bug #3176: osmocom debian packages are not install / upgrade tested added
Updated by osmith almost 4 years ago
- Related to Bug #3369: no automatic testing of Debian/Ubuntu packages added
Updated by osmith almost 4 years ago
- Related to Feature #4550: Integrate RPM spec files into Osmocom repositories and build them on OBS added
Updated by osmith almost 4 years ago
- move scripts that install all packages and try to run the services from docker-playground.git to osmo-ci.git. In docker-playground.git, we only have scripts for one image (nothing shared), so it fits better in osmo-ci.git
- create a new image centos-repo-install-test in docker-playground.git
- in osmo-ci.git:
- trigger building the images from docker-playground.git (there's already code for that)
- mount the shared script dir inside the containers
- run the scripts in the containers, with a distro parameter to do distro specific installation etc.
- adjust jenkins job builder yml to make use of the new structure and run it for centos, too
Updated by osmith almost 4 years ago
- % Done changed from 0 to 70
Patches submitted:
https://gerrit.osmocom.org/q/topic:centos-install-test
- https://jenkins.osmocom.org/jenkins/job/Osmocom-repo-install-centos/
- https://jenkins.osmocom.org/jenkins/job/Osmocom-repo-install-debian/
I just noticed that the workspace is not cleaned up properly yet:
java.nio.file.FileSystemException: /home/osmocom-build/jenkins/workspace/Osmocom-repo-install-debian/feed/nightly/label/osmocom-master-debian9/_repo_install_test_data/blacklist.txt: Operation not permitted
Furthermore, "Osmocom-repo-install-centos/feed=next" found a real error in the repository that needs to be fixed. This should not be problematic however, as it only occurs when installing osmo-trx-ipc and its debug symbols package (which AFAIK nobody is doing/trying at the moment):
Error: Transaction check error:
file /usr/lib/debug/usr/bin/ipc-driver-test-1.2.0.60.bb89-2.1.x86_64.debug conflicts between attempted installs of osmo-trx-ipc-debuginfo-1.2.0.60.bb89-2.1.x86_64 and osmo-trx-ipc-1.2.0.60.bb89-2.1.x86_64
Updated by osmith almost 4 years ago
osmith wrote:
I just noticed that the workspace is not cleaned up properly yet:
java.nio.file.FileSystemException: /home/osmocom-build/jenkins/workspace/Osmocom-repo-install-debian/feed/nightly/label/osmocom-master-debian9/_repo_install_test_data/blacklist.txt: Operation not permitted
Resolved with: https://gerrit.osmocom.org/c/osmo-ci/+/18518
Updated by osmith almost 4 years ago
- % Done changed from 70 to 90
osmith wrote:
Furthermore, "Osmocom-repo-install-centos/feed=next" found a real error in the repository that needs to be fixed. This should not be problematic however, as it only occurs when installing osmo-trx-ipc and its debug symbols package (which AFAIK nobody is doing/trying at the moment):
Error: Transaction check error:
file /usr/lib/debug/usr/bin/ipc-driver-test-1.2.0.60.bb89-2.1.x86_64.debug conflicts between attempted installs of osmo-trx-ipc-debuginfo-1.2.0.60.bb89-2.1.x86_64 and osmo-trx-ipc-1.2.0.60.bb89-2.1.x86_64
also fixed (in osmo-trx.git osmith/ipc branch, which is what the script for next is currently building).
Updated by osmith almost 4 years ago
rpm is failing on jenkins slave admin2-deb9build for some reason (is the machine broken?). But I can't look further into that now, I've configured the job to only run on build2-deb9build-ansible for now, this works fine.
error: rpmdb: BDB0060 PANIC: fatal region error detected; run recovery error: db5 error(-30973) from db->close: BDB0087 DB_RUNRECOVERY: Fatal error, run database recovery
Updated by laforge almost 4 years ago
On Wed, May 27, 2020 at 01:02:57PM +0000, osmith [REDMINE] wrote:
rpm is failing on jenkins slave admin2-deb9build for some reason (is the machine broken?).
the machine is fine, it runs tons of other services. There is plenty of disk space, and no errors
in the kernel log or other suspicious behavior.
- rebuilding the docker container
- restarting the deb9build lxc (can use "reboot" like a physical machine).
Updated by osmith almost 4 years ago
- Status changed from In Progress to Resolved
- % Done changed from 90 to 100
laforge wrote:
On Wed, May 27, 2020 at 01:02:57PM +0000, osmith [REDMINE] wrote:
I suggest:
- rebuilding the docker container
I've tried this yesterday, same result.
- restarting the deb9build lxc (can use "reboot" like a physical machine).
Debugging this now doesn't seem a good use of time, given that I also need to get the centos8 ttcn-3 implementation done. With the workaround of hardcoding a node, it is working now, and I'm certain that it is an issue with that specific jenkins node (admin2-deb9build).
I'll set this issue to resolved, as the centos packages install and systemd services test script is now rolled out.
Updated by laforge almost 4 years ago
On Thu, May 28, 2020 at 10:13:13AM +0000, osmith [REDMINE] wrote:
Debugging this now doesn't seem a good use of time, [...]
I'll set this issue to resolved, [...]
Then please at least create a separate ticket about the bug, as otherwise we
lose all knowledge of it.
Updated by osmith almost 4 years ago
laforge wrote:
On Thu, May 28, 2020 at 10:13:13AM +0000, osmith [REDMINE] wrote:
Debugging this now doesn't seem a good use of time, [...]
I'll set this issue to resolved, [...]Then please at least create a separate ticket about the bug, as otherwise we
lose all knowledge of it.
Updated by osmith almost 4 years ago
- Related to Bug #4567: rpm is failing on jenkins slave admin2-deb9build in docker added