Hi,
The 32 bit bit packages are located on the same directory as 64bits. I
tried adding a 32 bit repository in my project through direct link but it
requires an admin access so i could not complete my build.
Can I please ask the Admin to add 32bit support for AlmaLinux?
Regards,
James Dominic Guana
Hello,
in a package in home:dstoecker:base I do on an x86_64 system
osc build --local-package -j 4 --vm-type=qemu --no-verify openSUSE_Leap_15.0_ARM armv7l package.spec
that worked in the past, but now I get this:
[ 1s] setting hostarch to armv7l
[ 1s] booting qemu...
[ 1s] Using UART console
[ 1s] -nodefaults -no-reboot -nographic -vga none -runas qemu -net none -kernel /var/tmp/build-root/openSUSE_Leap_15.0_ARM-armv7l/.mount/boot/kernel -initrd /var/tmp/build-root/openSUSE_Leap_15.0_ARM-armv7l/.mount/boot/initrd -append root=/dev/sda rootfstype=ext3 rootflags=data=writeback,nobarrier,commit=150,noatime ext4.allow_unsupported=1 mitigations=off panic=1 quiet no-kvmclock elevator=noop nmi_watchdog=0 rw rd.driver.pre=binfmt_misc console= init=/.build/build -m 512 -drive file=/var/tmp/build-root/openSUSE_Leap_15.0_ARM-armv7l/img,format=raw,if=none,id=disk,cache=unsafe -device ,drive=disk,serial=0 -drive file=/var/tmp/build-root/openSUSE_Leap_15.0_ARM-armv7l/swap,format=raw,if=none,id=swap,cache=unsafe -device ,drive=swap,serial=1 -serial stdio -chardev socket,id=monitor,server,nowait,path=/var/tmp/build-root/openSUSE_Leap_15.0_ARM-armv7l/img.qemu/monitor -mon chardev=monitor,mode=readline -smp 4
[ 1s] /usr/lib/build/build-vm-qemu: line 210: -nodefaults: command not found
[ 1s] No buildstatus set, either the base system is broken (kernel/initrd/udev/glibc/bash/perl)
[ 1s] or the build host has a kernel or hardware problem...
It seems there is a command name missing in 4th line.
Does anybody know what goes wrong here?
Using openSUSE Tumbleweed with following packages installed
osc-0.176.0-1.1.noarch
qemu-6.2.0-41.1.x86_64
qemu-arm-6.2.0-807.4.x86_64
rpm-build-4.17.0-1.1.x86_64
Freedom in Peace
--
https://www.dstoecker.eu/ (PGP key available)
Hi folks.
On our OBS instance we introduced ubuntu:22.04 DoD project for x86_64 and
aarch64.
I took prjconf from https://build.opensuse.org/projects/Ubuntu:21.10/prjconf
and this worked quite well for one or two months.
Since 2022-03-08 in the morning the build results changed to "repository
broken" for all repos (standard, universe, universe-update, update), but only
for aarch64; x86_64 is still green.
From the dodup.log I could see that something changed on the repo/mirror:
2022-03-08 01:24:02: [1528] checking ubuntu:22.04/universe/aarch64...
updating metadata for deb repo at
https://in.mirror.coganng.com/ubuntu-ports/jammy/universe
...
2022-03-08 01:25:00: [1528] checking ubuntu:22.04/standard/aarch64...
updating metadata for deb repo at
https://in.mirror.coganng.com/ubuntu-ports/jammy/main
...
The repo/mirror itself seems to work correctly.
Looking at packages build against ubuntu:22.04 DoD project, I see complains
such as:
<error>unresolvable: preinstalls: nothing provides perl-modules-5.32</error>
I checked the repo and found perl v5.34, in prjconf I found the following
specifying concrete perl version:
...
Preinstall: sed gawk grep gzip debconf bash base-files base-passwd perl-modules-5.32 libperl5.32
...
Keep: m4 dpkg dpkg-dev libperl5.24 perl-modules-5.24 libdpkg-perl
Doing s/5.32/5.34/g didn't help, only breaking x86_64, too. Do you have any
tip/hint how to proceed here? Or where to look at for deeper debugging? Are
there any binary packages cached for DoD?
Thanks
--
Jan
Hey everyone,
As part of the beta program, you can now manage your beta features. This
allows you to disable a specific beta feature if somehow you don't want
to use it.
Details here:
https://openbuildservice.org/2022/03/17/beta-features-management/
Enjoy!
--
Dany Marcoux <dmarcoux(a)suse.de>
Full Stack Web Developer - Open Build Service
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5
90409 Nürnberg
Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Ivo Totev
> > Thus it seems like the DNS lookup did not work when the
> > scheduler was started. Simply restarting the schedulers
> > should fix this.
>
> The existing (broken) scheduled jobs also need to be removed.
>
> Which command removes the broken scheduled jobs?
They are files in the /srv/obs/jobs/<arch> directories,
just delete them with 'rm'.
And I was looking for an osc option/function. Silly me.
Thanks. After deleting those jobs and restarting the scheduler along with the workers it now builds again.
Regards
Mathias
Hi,
since a couple of days we have a problem with our obs workers.
We start the service
Mär 15 10:39:52 obsworker1 obsworker[2158]: Fetching initial worker code fromhttp://obs.uib.local:5252/getworkercode
Mär 15 10:39:52 obsworker1 systemd[1]: Finished OBS worker.
Mär 15 10:40:56 obsworker1 systemd[1]: Stopping OBS worker...
Mär 15 10:40:56 obsworker1 obsworker[2215]: Shutting down obsworker
Mär 15 10:40:56 obsworker1 obsworker[2224]: Name "BSConfig::reposervers" used only once: possible typo at ./bs_worker line 695.
Mär 15 10:40:56 obsworker1 obsworker[2226]: Name "BSConfig::reposervers" used only once: possible typo at ./bs_worker line 695.
Mär 15 10:40:56 obsworker1 obsworker[2225]: Name "BSConfig::reposervers" used only once: possible typo at ./bs_worker line 695.
Mär 15 10:41:14 obsworker1 obsworker[2215]: ..done
Mär 15 10:41:14 obsworker1 systemd[1]: obsworker.service: Succeeded.
But when I check the screen session I get
Name "BSConfig::reposervers" used only once: possible typo at ./bs_worker line 695.
worker started on port 33567 code ? build ?
Name "BSConfig::reposervers" used only once: possible typo at ./bs_worker line 695.
activating new worker code 390facb63336a0fa14d6dd84e2e15883
192.168.109.101 [2400]: rebooting...
Name "BSConfig::reposervers" used only once: possible typo at ./bs_worker line 695.
worker started on port 37155 code 390facb63336a0fa14d6dd84e2e15883 build d41d8cd98f00b204e9800998ecf8427e
fetching new buildcode f434116c0f239d7ec7be232d1928223a, mine was d41d8cd98f00b204e9800998ecf8427e
got job, run build...
2022-03-15 11:10:39: building 'zsync-curl' for project 'home:uibmz:opsi:4.2:stable' repository 'RHEL_8' arch 'x86_64'
fetching sources, getsources: connect to localhost:5352: Connection refused
build failed, marked as bad build host...
rpc failed: connect to localhost:5252: Connection refused
Here are the obs-server entries in syconfig
OBS_SRC_SERVER="192.168.109.101:5352"
OBS_REPO_SERVERS="192.168.109.101:5252"
Why does it try to connect to localhost?
obsworker1:/etc # rpm -qa | grep obs
obs-common-2.10.12-lp153.80.2.noarch
system-user-obsservicerun-2.10.12-lp153.80.2.noarch
system-user-obsrun-2.10.12-lp153.80.2.noarch
obs-worker-2.10.12-lp153.80.2.noarch
Regards
Mathias
--
opsi - Open PC-Server-Integration
das Open Source Client-Management-System von uib gmbh
Mathias Radtke
Entwicklung
eMail: m.radtke(a)uib.de
Tel. +49 6131 / 27561-0
Fax +49 6131 / 27561-22
uib gmbh
Bonifaziusplatz 1B
55118 Mainz
Internet:
https://uib.dehttps://opsi.org
Geschftsfhrer: Dr. Detlef Oertel, Erol Ülükmen
Handelsregister: Amtsgericht Mainz HRB 6942
On Tue, Mar 15, 2022 at 10:15:21AM +0000, Mathias Radtke wrote:
> OBS_SRC_SERVER="192.168.109.101:5352"
>
> OBS_REPO_SERVERS="192.168.109.101:5252"
>
>
>
> Why does it try to connect to localhost?
The source server to use is part of the job data. So the question
is why the scheduler has put localhost into the job data.
If your BSConfig.pm looks like the template we have, it looks
like:
my $hostname = Net::Domain::hostfqdn() || 'localhost';
our $srcserver = "http://$hostname:5352";
Thus it seems like the DNS lookup did not work when the
scheduler was started. Simply restarting the schedulers
should fix this.
So BSConfig.pm was like in the template.
However I changed the line now to
#my $hostname = Net::Domain::hostfqdn() || 'localhost';
my $hostname = 'obs.uib.local'
restarted the obs server to make sure all services will get the new hostname
restarted obsworker1 and I still get
2022-03-15 11:54:48: building 'opsiconfd' for project 'home:uibmz:opsi:4.2:development' repository 'RHEL_8' arch 'x86_64'
fetching sources, getsources: connect to localhost:5352: Connection refused
build failed, marked as bad build host...
rpc failed: connect to localhost:5252: Connection refused
Any suggestion where to look at?
Regards
Mathias
Hi,
It seems that the Prefer directive always prefer the first prefer but
ignore later prefers?
Since project conf is recursively inherited from repositories used, what
is the preferred way to override the Prefer defined in repositories by a
different one in my own project conf? Simply putting a Prefer line in my
project conf doesn't work as inherited projconfs always appear before my
projconf in the buildconfig.
Thanks,
Kai
Greetings!
I have a question regarding openSUSE build service command-line tool or osc.
I was trying to create a release workflow for a package so I experimented
with it in this repository https://github.com/uncomfyhalomacro/hellorust.
Does osc support environmental variables? Because I want that I can use
them as part of the `env` section on the config YAML file so that I can do
the following
```sh
OSC_USERNAME=yourusername OSC_PASSWORD=yourpassword osc co
home:yourusername/projectname
cd home:yourusername/projectname
# do some bunch of stuff here
OSC_USERNAME=yourusername OSC_PASSWORD=yourpassword osc vc -m "somemessage"
projectname.changes projectname.spec
OSC_USERNAME=yourusername OSC_PASSWORD=yourpassword osc commit
# done committing new release to devel project for projectname
```
Or maybe extend the `osc token --operations` functionality that includes
checkout, service, and commit?
I will be waiting for your reply.
Sincerely,
Soc Virnyl S. Estela
I have solved this issue. It needed to fully specify the path to the file at
-f option like this: %files lang -f %_sourcedir/%name-%version/%{name}.lang