Questions and Answers : Unix/Linux : Can't run LHC WU
Message board moderation

To post messages, you must log in.

AuthorMessage
River~~

Send message
Joined: 13 Jul 05
Posts: 456
Credit: 75,142
RAC: 0
Message 11347 - Posted: 14 Nov 2005, 9:20:10 UTC
Last modified: 14 Nov 2005, 9:34:01 UTC

hi I can't seem to get LHC to work with Linux.

Predictor, Einstein, and LHC all run OK on this windows box.

Predictor & Einstein work OK on this Linux box, but LHC does not.

Both boxes are similar hardware, each has twin Pentium III 665MHz processors. The linux box is running CentOS Linux and BOINC 5.2.7. BOINC downloaded LHC Sixtrack ver 4.66

There is no GUI support at all on the Linux box (no X, no KDE, etc) so I am driving BOINC using boinc_cmd, which seems to work OK for the other two projects.

Started it going last night, got up this morning and it has run off some Predictor WU ok, got half way through two Einsteins, but has downloaded and trashed over 60 LHC WU. Thank goodness for the wise coding that included a daily quota, or it would have been a lot more!

Might this be a problem with .so libraries? Does LHC expect differrent libraries to the other projects?

Is there something to do with the screensaver that might be different between LHC and the other two projects? As far as I know I am not trying to use a screensaver with any of the projects, but wondered if the absence of X might be relevant?

Any other ideas?

In the meantime I will set nomorework for LHC on the Linux box, no point trashing any more WU nor wasting my bandwidth downloading code & data that won't work.

River~~


ID: 11347 · Report as offensive     Reply Quote
River~~

Send message
Joined: 13 Jul 05
Posts: 456
Credit: 75,142
RAC: 0
Message 11348 - Posted: 14 Nov 2005, 12:47:51 UTC
Last modified: 14 Nov 2005, 12:52:11 UTC

This is an extract from the output from the BOINC client. It shows what happened to just one of the 69 WU, but as far as I can see they all produced the same error messages. The first and last two lines are the same, and it then repeats many times over - it would still be going round this loop if it had not eventually hit the quota.

Does the execv message at 22:23:20 offer a hint to anyone more knowledgeable than me?

R~~
---------

2005-11-13 21:06:08 [LHC@home] Sending scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi
2005-11-13 21:06:08 [LHC@home] Reason: To fetch work
2005-11-13 21:06:08 [LHC@home] Requesting 34560 seconds of new work, and reporting 1 results
2005-11-13 21:06:13 [LHC@home] Scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi succeeded
2005-11-13 21:06:14 [LHC@home] Started download of wnov1C_v6s4hvnom_mqx__11__64.305_59.325__10_12__6__30_1_sixvf_boinc103204.z
ip
2005-11-13 21:06:19 [LHC@home] Finished download of wnov1C_v6s4hvnom_mqx__11__64.305_59.325__10_12__6__30_1_sixvf_boinc103204.
zip
2005-11-13 21:06:19 [LHC@home] Throughput 9637 bytes/sec
2005-11-13 21:06:20 [---] request_reschedule_cpus: files downloaded
2005-11-13 22:23:20 [Einstein@Home] Pausing result l1_0359.5__0359.9_0.1_T12_S4lC_0 (left in memory)
2005-11-13 22:23:20 [LHC@home] execv(../../projects/lhcathome.cern.ch/sixtrack_4.66_i686-pc-linux-gnu) failed: error -1
2005-11-13 22:23:20 [LHC@home] Starting result wnov1C_v6s4hvnom_mqx__11__64.305_59.325__10_12__6__30_1_sixvf_boinc103204_0 usi
ng sixtrack version 466
2005-11-13 22:23:21 [LHC@home] Unrecoverable error for result wnov1C_v6s4hvnom_mqx__11__64.305_59.325__10_12__6__30_1_sixvf_bo
inc103204_0 (process exited with code 26 (0x1a))
2005-11-13 22:23:21 [LHC@home] Unrecoverable error for result wnov1C_v6s4hvnom_mqx__11__64.305_59.325__10_12__6__30_1_sixvf_bo
inc103204_0 (process exited with code 26 (0x1a))
2005-11-13 22:23:21 [---] request_reschedule_cpus: process exited
2005-11-13 22:23:21 [LHC@home] Computation for result wnov1C_v6s4hvnom_mqx__11__64.305_59.325__10_12__6__30_1_sixvf_boinc10320
4_0 finished
2005-11-13 22:23:21 [Einstein@Home] Resuming result l1_0359.5__0359.9_0.1_T12_S4lC_0 using einstein version 481
2005-11-13 22:24:21 [LHC@home] Sending scheduler request to http://lhcathome-sched1.cern.ch/scheduler/cgi
2005-11-13 22:24:21 [LHC@home] Reason: To fetch work
ID: 11348 · Report as offensive     Reply Quote
River~~

Send message
Joined: 13 Jul 05
Posts: 456
Credit: 75,142
RAC: 0
Message 11566 - Posted: 4 Dec 2005, 9:04:40 UTC

Now resolved, it worked when I came back to look at it again. I am not entirely sure of the problem but best guess is that it was a Linux file permissions issue.

If so the learning point from it is that it is safest to install and run the BOINC software always from the same username.

My suggestion would be to use a special username (boinc maybe) to run the BOINC software.
ID: 11566 · Report as offensive     Reply Quote
Gio27@LHC

Send message
Joined: 1 Jun 06
Posts: 3
Credit: 2
RAC: 0
Message 15023 - Posted: 8 Oct 2006, 19:06:00 UTC

Hi,
my name is Giovanni and I'm itaian.
I'm running linux and I'm having the same problem.

Can you tell me what you exactly do, for make LHC running?

Thank's

Gio


ID: 15023 · Report as offensive     Reply Quote

Questions and Answers : Unix/Linux : Can't run LHC WU


©2024 CERN