41) Message boards : ATLAS Application : ATLAS native 1.22 (Message 7741)
Posted 17 Aug 2022 by David Cameron
Post:
I have already Atlas-Native running and would be happy to help testing here with apptainer.

At the moment, there is singulary installed on my Ubuntu 20.04.4 LTS

Can you tell me please the exact instructions how I can install apptainer to the boxes?

Is Apptainer the same as CentOS..... ?

yeti


Hi Yeti,

Apptainer from CVMFS works on Ubuntu, at least on one of my machines with Ubuntu 21.10. So you should not have to install anything locally, just let the tasks use the version from CVMFS. The fallback to local singularity is only in case apptainer from CVMFS does not work.
42) Message boards : ATLAS Application : ATLAS native 1.22 (Message 7734)
Posted 16 Aug 2022 by David Cameron
Post:
I have also tested with Centos Stream 9 and it works fine with CVMFS and boinc installed from standard packages and apptainer from CVMFS: https://lhcathomedev.cern.ch/lhcathome-dev/result.php?resultid=3108671
43) Message boards : ATLAS Application : ATLAS native 1.22 (Message 7733)
Posted 15 Aug 2022 by David Cameron
Post:
This fixes the problem for one of my computers.
44) Message boards : ATLAS Application : ATLAS native 1.22 (Message 7732)
Posted 15 Aug 2022 by David Cameron
Post:
Version 1.22 attempts to fix the errors like "failed to create /var/lib/condor directory: mkdir /var/lib/condor: permission denied" which are seen in some situations with certain apptainer versions.

The change is to mount only the current working directory (eg /var/lib/boinc/slots/0) into the container rather than the top level directory (eg /var).
45) Message boards : ATLAS Application : ATLAS native 1.20 (Message 7726)
Posted 12 Aug 2022 by David Cameron
Post:
I made a silly mistake when releasing this version by forgetting to make the wrapper script executable so all tasks were failing. Version 1.21 fixes this.
46) Message boards : ATLAS Application : ATLAS native 1.20 (Message 7725)
Posted 11 Aug 2022 by David Cameron
Post:
ATLAS native 1.20 was just released which uses apptainer instead of singularity. At the moment apptainer functionality is identical to singularity. If apptainer does not work there is still a fallback to using singularity so tasks should work as normal for those who have a locally installed singularity. But this will be removed at some point in the future so we recommend that people who cannot use singularity/apptainer from CVMFS install a local version of apptainer instead of singularity. apptainer provides a backwards compatible "singularity" command so installing it will not break production tasks still relying on singularity.
47) Message boards : ATLAS Application : ATLAS vbox v.1.16 (Message 7713)
Posted 2 Aug 2022 by David Cameron
Post:
We have released v1.16 which contains an updated vboxwrapper version based on another pre-release of version 26205, so this one is "26205a".
48) Message boards : ATLAS Application : ATLAS vbox v.1.15 (Message 7712)
Posted 2 Aug 2022 by David Cameron
Post:
Not intention on my part, but a new kind of task with updated ATLAS simulation software was added recently to the set of tasks automatically submitted here. I've asked the experts to look into why these tasks fail.


The problem is the new software requires a library version not installed in the old VM we use here, so I have blocked these tasks from being submitted.
49) Message boards : ATLAS Application : ATLAS vbox v.1.15 (Message 7676)
Posted 29 Jul 2022 by David Cameron
Post:
Not intention on my part, but a new kind of task with updated ATLAS simulation software was added recently to the set of tasks automatically submitted here. I've asked the experts to look into why these tasks fail.

I have submitted manually a batch of 20 event tasks to keep the queue full.
50) Message boards : ATLAS Application : ATLAS vbox v.1.15 (Message 7668)
Posted 29 Jul 2022 by David Cameron
Post:
We just released v1.15 which uses a new vboxwrapper version 26205.
51) Message boards : ATLAS Application : ATLAS vbox v.1.14 (Message 7535)
Posted 5 Jul 2022 by David Cameron
Post:
I've made a new version 1.14 which adds the open_file tag to the vdi image as discussed in this thread.

This may increase stability for some people.
52) Message boards : ATLAS Application : ATLAS vbox v.1.13 (Message 7509)
Posted 4 Jul 2022 by David Cameron
Post:
I checked the full logs and there is a strange timeshift in the middle of the task:

2022-07-04 18:42:41,307 | INFO     | executing command: export ATLAS_LOCAL_ROOT_BASE=/cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase;source ${ATLAS_LOCAL_ROOT_BASE}/user/atlasLocalSetup.sh --quiet;source $AtlasSetup/scripts/asetup.sh AtlasOffline,21.0.15,notest --platform x86_64-slc6-gcc49-opt --makeflags='$MAKEFLAGS'
2022-07-04 16:43:00,181 | INFO     | nothing to verify (for remote files)


Did you change the system time or something like that? This could have confused the software into hanging until the time "caught up" with the previous time:

2022-07-04 18:40:12,136 | INFO     | will abort job monitoring soon since job state=finished (job is still in queue)
2022-07-04 18:41:13,716 | INFO     | monitor loop #110: job 0:5515971838 is in state 'finished'
2022-07-04 18:41:13,716 | INFO     | will abort job monitoring soon since job state=finished (job is still in queue)
2022-07-04 18:42:15,298 | INFO     | monitor loop #111: job 0:5515971838 is in state 'finished'
2022-07-04 18:42:15,298 | INFO     | will abort job monitoring soon since job state=finished (job is still in queue)
2022-07-04 18:42:50,074 | INFO     | finished stage-out for finished payload, adding job to finished_jobs queue
2022-07-04 18:42:52,785 | INFO     | job 5515971838 has state=finished


I checked some others' tasks and they look ok with a runtime of around one hour or less.
53) Message boards : ATLAS Application : ATLAS vbox v.1.13 (Message 7506)
Posted 4 Jul 2022 by David Cameron
Post:
I have submitted a bunch of 100 tasks with 20 events each, to generate some load.

If those finish too quickly I can generate more on demand.

Thank you, David.

Are these tasks scientifically useful or do we get them only for testing the new application?


The outputs of these tasks are not used for any science, but you can think of your contribution here as helping to improve the way we get science results out of the tasks on the production project.

Seems those 100 tasks went out quickly, so I put another 200 in.
54) Message boards : ATLAS Application : ATLAS vbox v.1.13 (Message 7503)
Posted 4 Jul 2022 by David Cameron
Post:
I have submitted a bunch of 100 tasks with 20 events each, to generate some load.

If those finish too quickly I can generate more on demand.
55) Message boards : ATLAS Application : ATLAS vbox v.1.13 (Message 7468)
Posted 30 Jun 2022 by David Cameron
Post:
I think this is because the split into parallel processes only happens after the first event is processed. At this point most of the necessary data and libraries are loaded into memory and so the sub-processes can then all share the memory. So if the task is only 2 events then there is only 1 event left to process in the parallel stage.
56) Message boards : ATLAS Application : ATLAS vbox v1.11 (Message 7463)
Posted 30 Jun 2022 by David Cameron
Post:
Please try again with version 1.13. There was a bug in the bootstrapping mechanism that may have caused your CVMFS issues that is now fixed.
57) Message boards : ATLAS Application : ATLAS vbox v.1.13 (Message 7462)
Posted 30 Jun 2022 by David Cameron
Post:
Version 1.13 is now available for Windows and Linux. It contains a fix in the bootstrapping that may help with CVMFS problems some people were having (see https://lhcathomedev.cern.ch/lhcathome-dev/forum_thread.php?id=569.
58) Message boards : ATLAS Application : ATLAS vbox v1.11 (Message 7436)
Posted 27 Jun 2022 by David Cameron
Post:
The reason for the download errors is that tasks get cancelled upstream if they are not completed within a certain time. This removes the input files from the server but sometimes the cancellation of the WU does not work so they are stay in the queue. I'll try to make the cancellation work better or alternatively set a shorter deadline for the tasks sent here.
59) Message boards : ATLAS Application : ATLAS vbox v1.11 (Message 7430)
Posted 24 Jun 2022 by David Cameron
Post:
I got a task and all files from app version 0.84.
Did you restart the BOINC server?

<edit>
Sorry, my fault!
You published only a Windows version.
</edit>


I've added a linux version now (1.12). It uses the same vdi as 1.11 for Windows.
60) Message boards : ATLAS Application : ATLAS vbox v1.11 (Message 7425)
Posted 24 Jun 2022 by David Cameron
Post:
This is the same as v1.10 but testing using a cloned vdi file. This is needed with the new multiattach mode which does not allow using the exact same image in multiple versions. See this thread on the CMS forum for more details: https://lhcathomedev.cern.ch/lhcathome-dev/forum_thread.php?id=562


Previous 20 · Next 20


©2024 CERN