frequent "compute errors"

AF
AF
Joined: 20 May 09
Posts: 1
Credit: 89775
RAC: 0
Topic 194394

I'm only having this difficulty (regularly) with Einstein@Home. Resetting the project usually works but the computer/application are frequently left unattended.

There is no one set of tasks or applications that this affects, as some of the same application work and some do not.

Suggestions?

Gundolf Jahn
Gundolf Jahn
Joined: 1 Mar 05
Posts: 1079
Credit: 341280
RAC: 0

frequent "compute errors"

The two tasks in your list that did error out show the "Can't acquire lockfile" message. You could do an Advanced search (over 3 or 6 months) for "acquire lockfile" to see threads covering that topic, using the link at the upper left corner of this page.

Sometimes it can be caused by using the BOINC CPU throttle mechanism, sometimes it's AV software locking the files, but I'm not sure if always a reason could be found.

Gruß,
Gundolf

Computer sind nicht alles im Leben. (Kleiner Scherz)

Keith
Keith
Joined: 11 Feb 05
Posts: 2
Credit: 1670431
RAC: 0

RE: I'm only having this

Quote:

I'm only having this difficulty (regularly) with Einstein@Home. Resetting the project usually works but the computer/application are frequently left unattended.

There is no one set of tasks or applications that this affects, as some of the same application work and some do not.

Suggestions?

Hi,

I think I am having a similar problem and I am beginning to think it there is a deadlock condition between two einsteing@home applications running on different processors on the same machine. I've been running einstein@home on two computers (both of which are on more or less permanently) for years now. In all honesty, I rarely look at detailed statistics because if I see general progress being made on the boinc statistics tab, then I am happy.

But today I just chanced to look in detail at my einsteing@home results for both computers and was somewhat disturbed to see the majority of work unites on both machines seem to fail (often after hours of cpu time) with a client error. See below for the error nformation but the gist seems to be that something is locking a file and various waits occur as retries are made to acquire a file lock and then eventually the whole thing times out.

Clearly, with all of these exit (0) I am receiving, Something seems to be locking a file einstein@home is trying to write to. I can only imagine that it is either my Norton 360 antivirus package or einstein@home itself. Both machines are running vista 32 bit on Intel core due processors. Both machines have two processors and I have been allowing up to 100% cpu on each processor, so i guess there is some possibility of a deadlock condition between two running instances of einsteing@home?

Anyway, for the moment I have restricted boinc so that it can only use one processor on each machine and I will see how that plays out. Any help or advice would be much appreciated though because this problem has resulted in the loss or waste of many weeks of cpu time which is very disappointing. I estimate that only one in three work packages succeeds.

Many thanks
Keith

core_client_version>6.6.36

too many exit(0)s

application '..\..\projects\einstein.phys.uwm.edu\einstein_S5R5_3.05_windows_intelx86_2.exe'.
Activated exception handling...
22:14:36 (7216): Can't acquire lockfile (32) - waiting 35s
22:15:11 (7216): Can't acquire lockfile (32) - exiting
22:15:11 (7216): Error: The process cannot access the file because it is being used by another process. (0x20)
2009-07-17 22:15:13.4570 [normal]: This program is published under the GNU General Public License, version 2
2009-07-17 22:15:13.4590 [normal]: For details see http://einstein.phys.uwm.edu/license.php
2009-07-17 22:15:13.4600 [normal]: This Einstein@home App was built at: Apr 10 2009 17:21:18

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5893653
RAC: 0

RE: I can only imagine that

Message 93396 in response to message 93395

Quote:
I can only imagine that it is either my Norton 360 antivirus package


Most probably, so make sure you set up your antivirus to NOT actively scan in the BOINC Data directory and its sub-directories. You can probably set somewhere in your AV scanner to omit or exclude directories. So add to it C:\ProgramData\BOINC\ which is the default BOINC Data directory path on Vista. Of course, if you changed the path to your Data directory, make sure to point at the correct one.

If you do want to scan the BOINC data directory, do so only after you exited BOINC.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.