Currently, on https://build.opensuse.org/project/show/devel:languages:misc,
all builds for OpenFactory:x86_64 and i586 are unresolveable.
It seems the on-going problem has something to do with
<path project="openSUSE:Factory" repository="snapshot"/>
in the meta configuration.
Is this mailing list to report this issue? or should I report it to bugzilla?
I have noticed some build service repositories, such as
have two directories with 15.3 in the name? Python only seems to have
Is this because we are part way through renaming everything and I had
not noticed the plan to do so? Or is it something else?
I’m trying to automate debin package build from GitHub repository.
All is working so well for "Debian 11 » (thanks BTW for this great workflow) that I decided to also apply the build to "Debian Unstable" and "Debian Testing » distribution.
It worked smoothly for "Debian Unstable », but for "Debian Testing » I’ve the message:
« Unresolved: have choice for bzr needed by obs-service-tar-scm: brz bzr"
and after searching (maybe bad research) I could not spot a proper explanation.
I read the code (https://github.com/openSUSE/obs-service-tar_scm) and could not find the potential issue.
I might not look at the right place... :3
URL of the project showing the issue:
_service file common to Debian 11/Debian Unstable and Debian Testing:
<service name="tar" mode="buildtime"/>
<service name="recompress" mode="buildtime">
<service name="set_version" mode="buildtime"/>
Any idea/hint ?
Thanks in advance for your help !
All the best
Attached a screenshot:
I ran into a strange build failure while doing a test compile of the
latest syslog-ng git snapshot:
[ 300s] syslog-ng-18.104.22.168.g628debb-1.1.ppc64le
[ 300s] error: unpacking of archive failed on file /usr/sbin/syslog-ng;6180fb04: cpio: open failed - No such file or directory
[ 300s] error: syslog-ng-22.214.171.124.g628debb-1.1.ppc64le: install failed
[ 300s] syslog-ng-python-126.96.36.199.g628debb-1.1.ppc64le
And the error is similar for all Factory / Tumbleweed variants. Older
distros work just fine. See:
Do you have any idea what could have gone wrong? I do not see anything
else suspicious in the build logs.
The Build Solutions Team has spent the last couple of weeks improving
the UI and the documentation of the workflows integration between OBS
You can find more details in our latest blog post:
The Build Solutions Team
Eduardo Navarro <enavarro(a)suse.com>
Open Build Service Frontend Developer
SUSE Software Solutions Germany GmbH
5 Nov 2021 07:11:14 Nicolas FORMICHELLA <nouloug2(a)gmail.com>:
A recent update to the Debian_Testing VM image/script triggered a rebuild on home:stig124:nnn/nnn, but it fails while creating the VM with a `mkswap` command not found error.
You can see the full build logs in the project.
While looking at boo, I noticed that the kernel and initrd of the VM
workers are different for different jobs.
For example, for SLES, qemu uses the kernel and initrd under
/var/cache/obs/worker/root_*/.mount/boot/kernel. But for CentOS, it uses
kernel under /boot. It looks like the SLES worker even runs the exact
kernel version of the target OS.
Could some one let me know how can I set my instance to run in such way?
we are trying to host our own instance of OBS 2.10 using SLES 15.2 as base system. But we are unable to build any project because the repo setup shows only as broken with - interconnect error: rpc timeout.
The production.log indicates that the proxy connection out of our LAN isn´t working, even though the credentials are deposited in BSConfig.pm and options.yml.
Error performing IssueTrackerUpdateIssuesJob (Job ID: f143a1hfgh67frgh-sd345fs00) from DelayedJob(issuetracking) in 131062.0ms: Errno::ETIMEDOUT (Failed to open TCP connection to bugzilla.opensuse.org:443 (Connection timed out - connect(2) for "bugzilla.opensuse.org" port 443)
Did anyone face similar problems hosting an OBS instance?
I would like osc to have command or plugin to behave similar to Fedora's koji build command. It is waiting until the running build of the server is finished (with any result) for all architectures. This is not very complicated and looks like something that I can implement on my own at holidays.
However, there is a design issue I would like to discuss before. Is there a better way to track the build status except polling /build/<project>/_result every few seconds or so? I found rabbit.opensuse.org server, but I am not sure if it is considered as a part of API and guaranteed to be here. Also AMQP support can be easily missed for the private OBS instances (like mine one).