Development for LHC@home

Hosts here are failing. Nothing to do but set NNW for the rest of the night.

All tasks failing after 7 minutes on the Windows 10 PC. Two still running on this Linux host.
Tullio

Thanks. The disk was full on the server. Note that this did not affect the beta app as it uses a different server. Over the next week or so all the services that are needed to support the Theory and CMS apps in the production project will be reviewed to ensure that they are production quality.

Sorry I didn't notice that last night, I was concentrating on the -beta. Is it OK now? My last failure was only 20 mins ago...

[Edit] Spoke too soon, and now I'm "over quota" too. [/Edit]

---

Final (?) hurdle overcome, jobs are starting to run again.

---

Started a host up by hand to check. Running OK now, thanks.
Laurence posted here what each console should eventually show but these logs seem somewhat confused:-

From the web server:-
MasterLog is OK
StartLog is OK
Starter.Log is OK
stderr.log shows cmsRun-stdout.log: No such file or directory.
stdout.log is OK.

No other logs are listed.

From the consoles:-
F1 is OK (boot& init)
F2 no output (should be running.log??)
F3 is OK (top)
F4 not sure, definitely a log of some sort. (Could be wrapper stdout,lots of CMSSW messages)
F5 shows cmsRun.stdout: No such file or directory. (should be wrapper stderr??)
F6 is OK (login)

cmsRun is using 80-90% CPU so presumably it's running OK.

You have a Condor job running on HostID 1033, currently showing 66 mins of "activity time" (whatever that means exactly...).

---

I have (hopefully) added some messages to the consoles to indicate what should be there even if they are blank.

Will hopefully now find the cmsRun-stdout.log.

Thanks, Ivan, but I've turned the host off again now. It will start itself up tonight. Not sure what the fate of that job will be but I don't expect it to resume successfully.

Will hopefully now find the cmsRun-stdout.log.

Yes, displayed in Console 2 and in the logs called: running.log

As you will.

---

What is the expected upload size for a job from the 250ev10ke batch?
(Logs+results)

What is the expected upload size for a job from the 250ev10ke batch?
(Logs+results)

The major root-result file for 250 events is about 66MB.

Thanks,CP.
I got about 80MB total, so with logs, it is in the ball-park.
Just a sanity check, as atlas is very different(for me).

What is the expected upload size for a job from the 250ev10ke batch?
(Logs+results)

The major root-result file for 250 events is about 66MB.

Yes, it varies +/- 10-20%. The logfile upload to the Condor server is the _condor_stdout, 130KB or so for a good job; the stderr that you see on the -dev website for your tasks is relatively small. I don't think anything else goes anywhere else, but I could be wrong.

---