High level of errors and invalids


Advanced search

Message boards : Problems and Help : High level of errors and invalids

Author Message
Profile Steve Hawker*
Send message
Joined: 6 Nov 12
Posts: 13
Credit: 98,551
RAC: 0
Message 13389 - Posted: 11 Nov 2014, 23:13:22 UTC
Last modified: 11 Nov 2014, 23:13:35 UTC

The errors fail within a few minutes so not a disaster. Would like to fix it though. From the std.err:

---------------------------
Exception caught: Worker application apparently died prematurely
Status: -9
---------------------------

Invalids run full term. I'm getting 2 invalids for every 1 valid. Much more of a disaster. There's nothing in the std.err that stands out but I'm far from an expert.

Any ideas on either please?

Profile Janus
Volunteer moderator
Project administrator
Avatar
Send message
Joined: 16 Jun 04
Posts: 4487
Credit: 2,094,806
RAC: 0
Message 13402 - Posted: 15 Nov 2014, 12:38:39 UTC
Last modified: 15 Nov 2014, 12:40:58 UTC

Quite likely it is an old version of a lib interfering somehow. There's a switch for BOINC (I can't remember exactly which) that allows you to pause it after each workunit before uploading the crash/success and before deleting the workunit slot directory. Use that and then look for a file called
in.crash.txt

It should have some pointers as to which library is messing up. Try to update that one and then give it another shot.
Some systems also put pointers to segfaults in the dmesg and /var/log/messages

(I'm assuming that you are using the "native" configuration option because the shipped libraries failed entirely?)

Profile Steve Hawker*
Send message
Joined: 6 Nov 12
Posts: 13
Credit: 98,551
RAC: 0
Message 13404 - Posted: 17 Nov 2014, 20:48:42 UTC - in response to Message 13402.
Last modified: 17 Nov 2014, 20:49:14 UTC

Quite likely it is an old version of a lib interfering somehow. There's a switch for BOINC (I can't remember exactly which) that allows you to pause it after each workunit before uploading the crash/success and before deleting the workunit slot directory. Use that and then look for a file called
in.crash.txt

It should have some pointers as to which library is messing up. Try to update that one and then give it another shot.
Some systems also put pointers to segfaults in the dmesg and /var/log/messages

OK, I'll give that a go

(I'm assuming that you are using the "native" configuration option because the shipped libraries failed entirely?)

Yes. I followed the instructions posted here: http://burp.renderfarming.net/forum_thread.php?id=2154&nowrap=true#12806

Profile vaughan
Send message
Joined: 12 Mar 05
Posts: 13
Credit: 2,794,995
RAC: 0
Message 13628 - Posted: 15 Mar 2015, 8:29:03 UTC

I get quite a large number of tasks that want to run for a very long time - like 500+ hours. I found it better to abort them. Why does this happen? Win 7 64 bit.
____________


Post to thread

Message boards : Problems and Help : High level of errors and invalids