Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 2
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 934 times and has 1 reply Next Thread
Rickjb
Veteran Cruncher
Australia
Joined: Sep 17, 2006
Post Count: 666
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
System Crash Left BOINC Comatose

Yesterday one of my computers had a system crash, and the filesystem containing the BOINC data directory was corrupted. Windows performed an automatic chkdsk on the volume, and rearranged things a bit, including putting 3 BOINC data files in a "found.000" directory.

When BOINC restarted, the boincmgr task list was empty, as was the messages list. Most menu options were greyed out. I could get a message saying that it (boincmgr) could not find a client (I forget how). boinc.exe was not in the task manager processes list.

I killed boingmgr and boinctray and then tried copying the one rescued orphan file that was not already there back into its slot directory, and started boincmgr again. The problem remained. I saved a snapshot of the BOINC data directory, then uninstalled and reinstalled BOINC. This left just a handful of new files in the BOINC data directory. I deleted these, then copied my snapshot files in. No go. I uninstalled and reinstalled BOINC a second time, and attached to WCG anew. Crunching for WCG was up and running again.

I was unhappy about losing the tasks in my work queue, expecting them to time out after their deadline. I need not have worried, as they have all been deemed "Detached", probably when I reattached the new BOINC installation to WCG. They have been re-sent, and some have been validated for their new owners already.

I've been around computers and WCG/BOINC for some time now, and found a way of dealing with this situation. However, less experienced or less motivated people may not have rescued it, and they may have dropped out of contributing to WCG. It wasn't a "user-friendly" situation, and I thought you should know.

BOINC 6.2.19 32-bit, Windows XP-64 SP2.

I haven't looked extensively at the BOINC log files, but stdoutdae.txt did not have anything useful. I still have the data snapshot, in case someone wants it.
The machine is overclocked, but it has never before crashed in this manner. There was a very brief mains power dropout (unusual) at about the time of the crash, and that may have been the trigger.
[Jul 6, 2010 3:51:16 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Sekerob
Ace Cruncher
Joined: Jul 24, 2005
Post Count: 20043
Status: Offline
Reply to this Post  Reply with Quote 
Re: System Crash Left BOINC Comatose

Amazing the dedication to rescue the tasks, Yes, rule number one going into BOINC/System recovery is staying off-line until happy things work again. BOINC has this tendency of telling the servers it's all gone, when reattaching and the client_state is void of any task tracks.

Power outs are not exactly unusual here, so the (Cheap) UPSs have long earned their keep. Coming home it's the answering machine that tells us... it switches itself on after a blip.
----------------------------------------
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All!
[Jul 6, 2010 4:14:45 PM]   Link   Report threatening or abusive post: please login first  Go to top 
[ Jump to Last Post ]
Post new Thread