Dr Watson has an episode whenever I crunch Einstein
18 Jul 2007 6:56:40 UTC
Topic 192973
(moderation:
I don't know if this has been talked about anywhere else..
but if it has can someone point me in the right direction. My keeps crashing when I crunch Einstein WUs because all my CPU is taken up by Dr Watson.
Physics is for gurls!
Dr Watson has an episode whenever I crunch Einstein
)
Well, your either going to have to unhide them or give us more info to work with, since this is a new one for me here on EAH.
Generally speaking though, I'm not understanding why Dr. Watson should be eating up any significant time or effecting EAH, unless it was either BOINC or the science app which was malfunctioning and triggering Watson in the first place.
Alinator
thanks..I have unhid my
)
thanks..I have unhid my computer. But i have set to no new work on Einstein to check that it is that that is setting Dr Watson off.
Physics is for gurls!
RE: thanks..I have unhid my
)
Somehow you've got SETI and Einstein mixed up. The Einstein task is tried to use SETI@Home optimized apps to run, who knows what that did to the project files. I would do a "reset project" from the Boinc Manager. That will delete all the files in the project directory and re-load the proper apps.
RE: RE: thanks..I have
)
Thanks very much ohiomike. I will try that :)
Physics is for gurls!
Hmmmm.... Well, the reset
)
Hmmmm....
Well, the reset might help, but there was only 1 out of the four failures which had the SAH anomaly.
I think it's more likely two stderr files ended up cross-linked, or possibly Scandisk ended up putting fragments of two stderr files together incorrectly after the crash on the reboot. It wasn't clear in your OP, but I'm going to assume you rebooted after Watson had it's 'psychotic episode'.
You might want to consider doing some disk maintenance and/or run some surface diagnostics to make sure the drive isn't getting 'tired' as well.
HTH,
Alinator
RE: Hmmmm.... Well, the
)
Yeah..resetting Einstein has not helped. Dr Watson is still having regular freakouts.
How can I tell if my drive is getting 'tired' and if it does get 'tired'..is there anything i can do about it?
Physics is for gurls!
RE: RE: You might want to
)
Your machine is listed as running Windows XP Home Edition. Assuming it is like Pro, you can get to a basic error check by selecting the drive itself (from Windows Explorer or My Computer),
Properties
Tools
Error Checking
Check Now...
and check the boxes for
Automatically fix file system errors
Scan for and attempt recovery of bad sectors
hit start, and expect to wait a few hours (depending on disk size and speed). Most likely it will find an application in conflict, and ask your permission to seize the disk to its sole use on your next reboot. I'd suggest agreeing, the rebooting as soon as you can afford to do without the machine for a while.
You can get lots more data on recent "heavy breathing" using an application which reviews the SMART data, but interpreting that is not for the faint-hearted (it shows lots of loggable episodes on new and excellent drives). By contrast the scan I've suggested will show if actual errors are found on your drive, not just cases where it had to retry or such.
I suggest disabling the Dr.
)
I suggest disabling the Dr. Watson program as it is useless IMHO to the average computer user. Unless you have more than the occasional inherent Windows hangup or repeated problems with Windows and/or applications crashing then I don't believe the program has any merit. And it seems from your initial post that [url=http://en.wikipedia.org/wiki/Dr._Watson_(debugger)]Dr. Watson[/url] is the cause and not a solution as it should be. The program can be easily disabled and re-enabled to see if you notice anything going awry beyond it. gl
RE: Yeah..resetting
)
As a first pass you could run Scandisk with the surface scan option checked. That should help find any really 'weak' clusters on the drive. If you wanted to get more thorough, there are other drive diagnostics available which 'dig' deeper and are more comprehensive. Keep in mind with the more elaborate the testing and the bigger the drive/partition is, it can take a loooong time to fully analyize it.
One thing I find somewhat curious here is your host appears to start out fine on a run but then later fails, apparently when it tries to go read in the checkpoint file on a restart. Perhaps another thing to try would be to exclude the BOINC folder and it sub-folders from AV scanning, just in order to rule out a BOINC file read timeout as the problem.
Regarding the suggestion about Watson, that certainly is a possibility, but my experience is that Watson doesn't trigger on it's own out of the blue. Something really bad has to happen, usually with something to do with the Win32 API or worse for it to come into the picture. Also you might be able to get some more insight into what's going on by looking over Windows event logs around the time of the failure for clues.
Alinator
RE: Regarding the
)
I concede that Alinator is right on the Dr. Watson issue concerning my suggestion. (The main box I use is pretty minimalistic and streamlined without what I consider bloat (most normal startup programs, etc., etc.)
Here is an article that correlates with Alinator's advice on looking over the event logs.