One host with constantly restarting tasks


Advanced search

Message boards : Problems and Help : One host with constantly restarting tasks

Author Message
Profile noderaser
Project donor
Avatar
Send message
Joined: 28 Mar 06
Posts: 512
Credit: 1,553,018
RAC: 79
Message 13008 - Posted: 10 Jul 2014, 3:04:21 UTC

I've noticed that one of my hosts 59130, seems to get some tasks every now and then that just stall out. I'll notice that through the tasks list here on the BURP site (when I get a "timed out" or the like) and check on the BOINC manager, only to discover that the task in question will restart at 0 time when BOINC manager is opened. It's not all of the tasks all of the time, and it doesn't seem to be any particular session that is triggering these errors. This is a new behavior that has just started over the last month or so. The host in question is my DVR, which has been a reliable cruncher up to this point; it's a Win7 Phenom x3 with 6 GB of RAM.

Upon further reflection, it's possible that this started with the "Blender v4.88 (mt)" app, as I don't recall there being any of this type of error before its release in May.

Here's an excerpt which stands out from the log on the latest task caught in a restart loop, which I manually aborted. It would appear that this bit appears each time the task is "restarted". Task 8099657

|('Observer claims to be done with all the scenes',)


---------------------------
Exception caught: BOINC kindly asks us to exit
Status: 0
---------------------------
terminate called after throwing an instance of 'Exception'

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
boinc_init_diagnostics() completed
boinc_init_options() completed
boinc_get_init_data() completed
CPU performance profile completed: 1744348976.073574 fpops, 4971870339.570176 iops reported. p_c is 1404901397.427399
Mapping logical files to physical destinations:
in => in
out.zip => ../../projects/burp.renderfarming.net/ses0000002243frm0000002940prt00001_0_0
./windows_zip.exe => ./windows_zip.exe
./windows_unzip.exe => ./windows_unzip.exe
Project Directory Base => C:\ProgramData\BOINC/projects/burp.renderfarming.net
____________

funkydude
Send message
Joined: 23 Dec 13
Posts: 275
Credit: 2,478,281
RAC: 0
Message 13009 - Posted: 10 Jul 2014, 8:54:09 UTC - in response to Message 13008.

Try resetting the project on the affected machine.

Profile Janus
Volunteer moderator
Project administrator
Avatar
Send message
Joined: 16 Jun 04
Posts: 4487
Credit: 2,094,806
RAC: 0
Message 13011 - Posted: 10 Jul 2014, 19:41:12 UTC - in response to Message 13008.
Last modified: 10 Jul 2014, 19:45:53 UTC

Exception caught: BOINC kindly asks us to exit


This is mostly always a settings thing ("kindly" is the keyword - had it said "forced us to abort right now" it would have been something else). Something in the preferences is causing BOINC to stop a task early. Check any option that has to do with whether jobs are allowed to run to completion: memory, swap, switch between tasks should be very high, suspend to memory rather than abort should be on (!) and so on etc. etc.
Looks like it happens every half an hour or every 15 mins or so - is something else sometimes running at those intervals?

Profile noderaser
Project donor
Avatar
Send message
Joined: 28 Mar 06
Posts: 512
Credit: 1,553,018
RAC: 79
Message 13040 - Posted: 17 Jul 2014, 1:25:34 UTC
Last modified: 17 Jul 2014, 1:26:19 UTC

Hmm, looks like the host had been getting "bad" preferences from somewhere, that specified 10% memory usage while idle, and 60% when active... Which might explain why the tasks seemed to restart when the computer was used, i.e. to look at why a task was stalling. I updated the preferences at BAM and made sure they transferred properly to the host, will wait and see if that was indeed the problem.
____________


Post to thread

Message boards : Problems and Help : One host with constantly restarting tasks