I'm getting an error with LHC ATLAS jobs, seems like a `tar` command is failing when it is trying to assign ownership of files it is unarchiving during the job setup. Anyone know what might be causing this?

Running on Fedora Silverblue 36. Using BOINC Container and CVMFS Container.

<core_client_version>7.16.17</core_client_version>
<![CDATA[
<message>
process exited with code 195 (0xc3, -61)</message>
<stderr_txt>
00:44:11 (10066): wrapper (7.7.26015): starting
00:44:11 (10066): wrapper: running run_atlas (--nthreads 12)
01:08:55 (142230): wrapper (7.7.26015): starting
01:08:55 (142230): wrapper: running run_atlas (--nthreads 12)
[2022-11-05 01:09:14] Arguments: --nthreads 12
[2022-11-05 01:09:14] Threads: 12
[2022-11-05 01:09:14] Checking for CVMFS
[2022-11-05 01:09:14] No cvmfs_config command found, will try listing directly
[2022-11-05 01:09:14] CVMFS is ok
[2022-11-05 01:09:14] Efficiency of ATLAS tasks can be improved by the following measure(s):
[2022-11-05 01:09:14] The CVMFS client on this computer should be configured to use Cloudflare's openhtc.io.
[2022-11-05 01:09:14] Further information can be found at the LHC@home message board.
[2022-11-05 01:09:14] Using apptainer image /cvmfs/atlas.cern.ch/repo/containers/fs/singularity/x86_64-centos7
[2022-11-05 01:09:14] Checking for apptainer binary...
[2022-11-05 01:09:14] apptainer is not installed, using version from CVMFS
[2022-11-05 01:09:14] Checking apptainer works with /cvmfs/atlas.cern.ch/repo/containers/sw/apptainer/x86_64-el7/current/bin/apptainer exec -B /cvmfs /cvmfs/atlas.cern.ch/repo/containers/fs/singularity/x86_64-centos7 hostname
[2022-11-05 01:09:14] fedora-pc
[2022-11-05 01:09:14] apptainer works
[2022-11-05 01:09:14] Set ATHENA_PROC_NUMBER=12
[2022-11-05 01:09:14] Starting ATLAS job with PandaID=5652539918
[2022-11-05 01:09:14] Running command: /cvmfs/atlas.cern.ch/repo/containers/sw/apptainer/x86_64-el7/current/bin/apptainer exec -B /cvmfs,/var/lib/boinc/slots/0 /cvmfs/atlas.cern.ch/repo/containers/fs/singularity/x86_64-centos7 sh start_atlas.sh
[2022-11-05 01:09:14] Job failed
[2022-11-05 01:09:14] ++ pwd
[2022-11-05 01:09:14] + export RUNTIME_CONFIG_DIR=/var/lib/boinc/slots/0/
[2022-11-05 01:09:14] + RUNTIME_CONFIG_DIR=/var/lib/boinc/slots/0/
[2022-11-05 01:09:14] + mv ATLAS.root_0 EVNT.30559922._003570.pool.root.1
[2022-11-05 01:09:14] + tar --strip-components=5 -xvf input.tar.gz
[2022-11-05 01:09:14] tar: pandaJobData.out: Cannot change ownership to uid 4871, gid 1028: Invalid argument
[2022-11-05 01:09:14] tar: runpilot2-wrapper.sh: Cannot change ownership to uid 4871, gid 1028: Invalid argument
[2022-11-05 01:09:14] tar: Exiting with failure status due to previous errors

I have only run Atlas tasks using VB though that in itself is not straightforward on Linux for anyone without a fair degree of competence using the operating system and tracking down errors.

My experience is that questions posed on the LHC boards usually get responses fairly quickly so worth posting there as well if you haven't already done so.

Thanks for the reply. I posted on the LHC board already and haven’t gotten any response for a few days. It does seem that the LHC jobs are a bit more complicated to get working than the typical BOINC projects.

No answer does not mean you will be ignored.

Native tasks at LHC@home are complex and not easy to set up.
That's true.
Hence, keep the setup as simple as possible and follow the howtos there.

You didn't get an answer so far because nobody, even the experts could deal with the additional complexity you want to add.
Other people might not be available atm for different reasons like busy with other work, on vacation, being ill ...