Occasional Errors

Message boards : Number crunching : Occasional Errors

To post messages, you must log in.

AuthorMessage
Virtual Boss*
Avatar

Send message
Joined: 10 May 08
Posts: 35
Credit: 713,981
RAC: 0
Message 61294 - Posted: 21 May 2009, 11:26:39 UTC

One of my crunchers has had a few errors.

A window pops up stating that "The application was requested to terminate in an unusual way"


Stderror txt contains:

<message>
The system cannot find the path specified. (0x3) - exit code 3 (0x3)
</message>

An example

Any thoughts?
ID: 61294 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Hammeh

Send message
Joined: 11 Nov 08
Posts: 63
Credit: 211,283
RAC: 0
Message 61296 - Posted: 21 May 2009, 12:10:37 UTC

Seems like the checkpoint file can't be located. Have you tried a fresh reinstall and install of the latest version of the BOINC client?
ID: 61296 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Virtual Boss*
Avatar

Send message
Joined: 10 May 08
Posts: 35
Credit: 713,981
RAC: 0
Message 61297 - Posted: 21 May 2009, 13:56:12 UTC - in response to Message 61296.  

Hammeh,

Thanks for your response, you are probably correct in your analysis.

I too was thinking it was a failure to access a file but couldn't work out which one, or more importantly why this only happens occaisonally and only recently.

I have been poring through various logs etc and I think I have found the answer.

Each failure was preceeded by an entry in the event viewer system log stating that virtual memory was too low, and was being increased. Obviously this caused a lack of resources during an attempt to checkpoint.

I then realized that the second fastest of my three paging files had somehow vanished.
The fastest is fully allocated, but the slowest on C drive has only 2MB of 1000 allocated at startup. When the first is fully utilised Windows then claims more space on the C drive allocation, causing a temporary resource shortage.

This was due to a CPDN HADAM3P being suspended and a Rosetta task running, both using a lot of VMEM (total more than paging file #1).

I will reinstate my missing paging file and monitor, and am confident that this will resolve the issue.
ID: 61297 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : Occasional Errors



©2024 University of Washington
https://www.bakerlab.org