climateprediction.net (CPDN) home page
Thread 'New Problem'

Thread 'New Problem'

Message boards : Number crunching : New Problem
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user760
Avatar

Send message
Joined: 10 Aug 04
Posts: 94
Credit: 309,849
RAC: 0
Message 5496 - Posted: 20 Oct 2004, 0:00:34 UTC
Last modified: 20 Oct 2004, 0:01:12 UTC

This is the third time this has happened. When I have the client suspended and reboot my computer (3.4G HT XP) the current WU's are lost and I get two new WU's and start crunching from scratch. Anyone else experience this?

<img><a> Team Site Link</a>"The world is a progressively realized community of interpretation."
ID: 5496 · Report as offensive     Reply Quote
ChinookFoehn

Send message
Joined: 7 Aug 04
Posts: 83
Credit: 410,895
RAC: 0
Message 5497 - Posted: 20 Oct 2004, 0:11:33 UTC - in response to Message 5496.  
Last modified: 18 Dec 2004, 8:20:18 UTC

ID: 5497 · Report as offensive     Reply Quote
ProfileThyme Lawn
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1283
Credit: 15,824,334
RAC: 0
Message 5506 - Posted: 20 Oct 2004, 7:21:46 UTC - in response to Message 5496.  

&gt; This is the third time this has happened. When I have the client suspended
&gt; and reboot my computer (3.4G HT XP) the current WU's are lost and I get two
&gt; new WU's and start crunching from scratch. Anyone else experience this?

Are you still on 4.09 after your bad experience with 4.13 Dogbytes? If so you should exit BOINC before rebooting, otherwise you risk getting an unrecoverable client_state.xml corruption (4.13 detects the corruption and automatically reverts back to client_state_prev.xml).
<br><a href="http://www.teampicard.net"><img src="http://www.teampicard.net/templates/fisubice/images/phpbb2_logo.jpg"></a><a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3">Join us here</a>
ID: 5506 · Report as offensive     Reply Quote
Dave Peachey

Send message
Joined: 5 Aug 04
Posts: 11
Credit: 2,356,953
RAC: 0
Message 5547 - Posted: 21 Oct 2004, 21:24:02 UTC
Last modified: 21 Oct 2004, 21:54:48 UTC

Guys,

I'm getting similar problems - it started with BOINC v4.05 and now continues with v4.09 (I haven't found, and don't intend to install, v4.13 until it's on the CPDN download page). The machine is an un-overclocked dual Athlon MP2800 system running Windows XP and, in essence, I've had two slightly different things happen - each with equally disasterous results:

1) Firstly with v4.05 BOINC client - last weekend, having got back from several week's holiday, I found that one of the two WUs had completed (the other was 99+% complete) but hadn't uploaded because my i/net connection had switched off in my absence. I reset the i/net connection, did a manual update, got the system to trickle and download two new WUs to replace the one complete and one nearly complete WU. However, as the completed WU didn't seem to want to upload, I suspended and then closed the BOINC application, rebooted the machine and restarted BOINC - only for the system to lose all references to my attachment to the CPDN project, clear out all WUs (completed, partially completed and both new ones)!

So, I uninstalled BOINC v4.05, cleared out the CPDN directory and installed BOINC v4.09, reattached to CPDN and then downloaded two new WUs. So far so good (hah bloody hah!), you might think ...

2) This evening, I did a couple of o/s critical updates and had to reboot the machine so, as per usual, I suspended and then closed the client before rebooting. However, on restarting the BOINC GUI executable had gone missing (as in, has been deleted!). Having no other obvious course of action, I reinstalled the v4.09 BOINC application but, for some unknown reason, the system chose to malfunction and upload the the existing partially completed WUs with client error messages.

Yes, I did manage to download another pair of WUs but these two malfunctions in less than a week have cost me six WUs - two complete or as near as damnit (forty or more days processing lost forever), two unstarted WUs and two partially completed WUs uploaded with client errors.

To say that I'm pissed off would be an understatement! I _presume_ due to a lack of any previous troubles with the machine when running "classic" CPDN that the BOINC client is the problem area. I can't say this is leaving me overly impressed!!

This puts my sorry statistics for my three machines since I started CPDN on BOINC on 5th August at:
23 = downloaded WUs
2 = uploaded results (completed successfully)
8 = aborted (uploaded with "client error" results)
5 = lost forever
8 = still processing
Not a good record so far - for me or for BOINC!

Dave
ID: 5547 · Report as offensive     Reply Quote

Message boards : Number crunching : New Problem

©2024 cpdn.org