| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Help Stop TB
Thread: Investigating Increased Error Rate in Help Stop TB |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 26
|
|
| Author |
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges: 
|
We have identified the issue and are working on finding and correcting the cause. Curious as to what the issue actually is, as the last Beta batch from July 1st worked just fine, even on 32bit Windows Server 2003/Windows XP hosts, but there seem to be a lot of those (random?) memory issues with regular WUs since right after the Beta run ended... Thanks, armstrdj  Ralf |
||
|
|
Gurra
Cruncher Joined: Sep 11, 2006 Post Count: 33 Status: Offline Project Badges: 
|
One Windows machine just got 5 WUs, all different batches but all with sequence #3, they all errored out straight away with (unknown error) - exit code -1073741819 (0xc0000005)
----------------------------------------Another Windows machine got 5 WUs, mostly different batches and with sequences #3 and #4, all bar one errored out with the above error, the last one with inconsistent extended attributes. A third Windows machine got a single WU - batch 4975 sequence 4 that is sitting in pending verification on two Darwin machines and one Windows machine. It appears to be processing normally  [Edit 1 times, last edit by Gurra at Jul 22, 2016 3:16:53 PM] |
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: 
|
We will begin releasing new work again on Help Stop TB shortly. We have tracked down the issue to the latest update done to the Windows 64 bit build of the application. We are reverting back to the previous Windows 64 bit version, version 7.14.
Thanks, armstrdj |
||
|
|
widdershins
Veteran Cruncher Scotland Joined: Apr 30, 2007 Post Count: 677 Status: Offline Project Badges:
|
I don't understand how that could be the root issue. I and others were seeing these memory errors on linux, would they be affected by the Windows build?
|
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges:
|
For Help Stop TB each workunit is a time slice of a simulation and it relies on the output of the previous slice to run. The windows build was producing bad output and when it was fed back into the system those runs saw the increased memory usage. The next slice could run on any platform regardless of where the previous slice ran, which is why you saw them on Linux.
----------------------------------------Thanks, armstrdj [Edit 1 times, last edit by armstrdj at Jul 22, 2016 8:03:25 PM] |
||
|
|
widdershins
Veteran Cruncher Scotland Joined: Apr 30, 2007 Post Count: 677 Status: Offline Project Badges:
|
Ah, I understand now, thanks for the explanation.
Will you be able to wind back the affected simulations to the last known good output slice just before the Windows runs corrupted things rather than start each simulation from the beginning? |
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges:
|
Yes the simulations affected will be restarted at the last good input.
Thanks, armstrdj |
||
|
|
tmedve
Senior Cruncher USA Joined: Nov 16, 2004 Post Count: 191 Status: Offline Project Badges: 
|
I've got 8 tasks with 2 running from this fixed batch (I think). 20 minutes in and all seems well.
|
||
|
|
widdershins
Veteran Cruncher Scotland Joined: Apr 30, 2007 Post Count: 677 Status: Offline Project Badges:
|
Yes the simulations affected will be restarted at the last good input. That's great news. It's good to know that all the previous work on these simulations won't be in vain. |
||
|
|
Sandvika
Advanced Cruncher United Kingdom Joined: Apr 27, 2007 Post Count: 112 Status: Offline Project Badges:
|
I'm delighted to hear that HST will soon be back in business. I've reverted to OET and am 6 real days away from completing 20 crunching years, so will stick with it until it is done and hope HST is ready again by then.
----------------------------------------  |
||
|
|
|