Development for LHC@home

11/23/16 15:39:11 (pid:4115) DockerProc::Detect()
11/23/16 15:39:11 (pid:4115) DOCKER is undefined.
11/23/16 15:39:11 (pid:4115) DockerAPI::detect() failed to detect the Docker version; assuming absent.


Starter.log..

Task hangs at stderr.txt --> " HTCondor ping --->0"

Tested several times---ideas?

There are no jobs due to a problem with our HTCondor servers. Things are running well but it was running out of memory due to the number of jobs. Tomorrow we will add another bigger machine.

Thanks for the info.
Much appreciated.

Some observations:

It uses very little memory; about 1 GB for a 4 core task.Disk space about 0.75GB per task plus 450 MB for the image file.

However, the startup/shutdown of each job takes about 2.5 minutes.
Compared to the run-time of about 15 to 20 min (in my case)it seems very wasteful.

These jobs are too short to run efficiently on multi-core tasks.

Will see, how they behave as single core tasks.

"Finished_x.log" and "running.log" does not contain any info.(just dummy argument)

BTW the ...

Guest Log: [INFO] Job finished in slot2 with unknown exit code.

in the stderr.txt is not particularly helpful, either.

Efficiency is very bad on single core tasks as well.

It is about 66%. For about 1/3 of the job duration the CPU is below 5%.

There is no significant disk or network activity, either.

I did a short test. 2 jobs done on single core VM.
Run time 57 min 51 sec
CPU time 47 min 14 sec

http://lhcathomedev.cern.ch/vLHCathome-dev/result.php?resultid=288738

I noticed that the process 'bc' is running several times during 1 job.

These are just test jobs to exercise the system. ALICE is currently busy with the p-Pb heavy ion run so will not be able to make any progress until next year.

Boinc 7.7.2 - Virtualbox 5.1.22

https://lhcathomedev.cern.ch/lhcathome-dev/result.php?resultid=324669

maxCPUs=1 and maxJobs=1

206 (0x000000CE) EXIT_INIT_FAILURE

Broken after 11 min. duration-time. RDP was shown. (console)

Alice will be taking a rest for a while.

I have a large number of OLD Alice, Atlas , LHCb , Benchark, etc., results still listed under my Tasks. Some of these are over 2 years old so surely can't still be useful. Is it not about time for a purge of these to free up some server space? I know some purges in the past haven't gone well and been over agressive but perhaps a 6 months limit (?) might allow enough records for task comparison by the user and remove redundant tasks.

Laurence will be back around August 10

---

Mad Scientist For Life