Message boards : CMS Application : New Version v47.30
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Laurence
Project administrator
Project developer
Project tester
Avatar

Send message
Joined: 12 Sep 14
Posts: 1021
Credit: 274,753
RAC: 0
Message 3999 - Posted: 8 Aug 2016, 7:51:44 UTC

This new version provides HTCondor 8.4.8 using the standard CernVM production distribution. It should finally solve the suspend/resume problems and is half the size of the previous image. Multi-core has also been enabled.
ID: 3999 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Crystal Pellet
Volunteer tester

Send message
Joined: 13 Feb 15
Posts: 1010
Credit: 591,653
RAC: 0
Message 4000 - Posted: 8 Aug 2016, 9:18:52 UTC
Last modified: 8 Aug 2016, 10:14:06 UTC

I think it's a bad idea to run 8 CMS-jobs in one 2048 MB Virtual Machine.

The task didn't succeed:

2016-08-08 11:52:54 (7128): Guest Log: [INFO] CMS application starting. Check log files.
2016-08-08 11:59:15 (7128): Guest Log: [ERROR] No jobs were available to run.
2016-08-08 11:59:15 (7128): Guest Log: [INFO] Shutting Down.
2016-08-08 11:59:15 (7128): VM Completion File Detected.
2016-08-08 11:59:15 (7128): VM Completion Message: No jobs were available to run.


Multicore VM's do not get jobs or just coincidence.
2 single core CMS-VM's are running cmsRun.
ID: 4000 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Laurence
Project administrator
Project developer
Project tester
Avatar

Send message
Joined: 12 Sep 14
Posts: 1021
Credit: 274,753
RAC: 0
Message 4002 - Posted: 8 Aug 2016, 13:06:15 UTC - in response to Message 4000.  

The memory scaling configuration hasn't been implemented. It was tried with Theory previously but was pulled as there was a problem. It also looks like it can't be done per application.
ID: 4002 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Rasputin42
Volunteer tester

Send message
Joined: 16 Aug 15
Posts: 966
Credit: 1,201,500
RAC: 0
Message 4004 - Posted: 8 Aug 2016, 14:20:52 UTC
Last modified: 8 Aug 2016, 14:26:20 UTC

I just did a project reset, but i am getting a 47.20 vbox 64 task.

EDIT> and it is the same .vdi file as before.
ID: 4004 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
computezrmle
Avatar

Send message
Joined: 28 Jul 16
Posts: 263
Credit: 232,222
RAC: 0
Message 4061 - Posted: 12 Aug 2016, 8:01:49 UTC

The CMS app fails if the VM runs on more than 1 core and the standard setting of 2048 MB RAM.
A VM running on 2 cores (configured via preferences page) and 3300 MB RAM (configured via app_config.xml) started fine and is currently running 2 jobs.

The RAM settings are lend from ATLAS and may be to high for CMS.
What are the minimum RAM requirements for CMS?
ID: 4061 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile ivan
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar

Send message
Joined: 20 Jan 15
Posts: 1093
Credit: 6,893,316
RAC: 0
Message 4062 - Posted: 12 Aug 2016, 8:13:16 UTC - in response to Message 4061.  

What are the minimum RAM requirements for CMS?

I think for GRID work we currently quote 1.5 GiB/core. For some analysis tasks we need more than that, but for the pure Monte-Carlo stuff that we have to run here (can't access stored data to do analysis) that should be more than adequate.
ID: 4062 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Rasputin42
Volunteer tester

Send message
Joined: 16 Aug 15
Posts: 966
Credit: 1,201,500
RAC: 0
Message 4063 - Posted: 12 Aug 2016, 8:31:17 UTC
Last modified: 12 Aug 2016, 8:33:27 UTC

I had a four core task fail a 5GB,but works at 5248MB (5GB+128MB)

I think, it is important to know exactly, how much memory is needed.
ID: 4063 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile ivan
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar

Send message
Joined: 20 Jan 15
Posts: 1093
Credit: 6,893,316
RAC: 0
Message 4071 - Posted: 15 Aug 2016, 11:57:33 UTC - in response to Message 3999.  

This new version provides HTCondor 8.4.8 using the standard CernVM production distribution. It should finally solve the suspend/resume problems and is half the size of the previous image. Multi-core has also been enabled.

What's the minimum version of BOINC required for this?
ID: 4071 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile ivan
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar

Send message
Joined: 20 Jan 15
Posts: 1093
Credit: 6,893,316
RAC: 0
Message 4072 - Posted: 15 Aug 2016, 22:59:30 UTC - in response to Message 4071.  

This new version provides HTCondor 8.4.8 using the standard CernVM production distribution. It should finally solve the suspend/resume problems and is half the size of the previous image. Multi-core has also been enabled.

What's the minimum version of BOINC required for this?

I had a Linux box on 7.2.33 that was having kernel panic trying to run the multicore tasks. I finally got it upgraded to 7.2.42 but it still kernel-panics. My other Linux machines also fail on CMS and Benchmark. I've shut them down.
ID: 4072 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : CMS Application : New Version v47.30


©2020 CERN