Re-download or Terminate work in process when hard drive crashed?
2 Oct 2005 23:26:43 UTC
Topic 189962
(moderation:
I had a hard drive crash that resulted in installing a new drive and doing a clean install. Two work units that show "in process" are not running on my computer. Is there a way to re-download or terminate these jobs to clean up the mess?
Re-download or Terminate work in process when hard drive crashed
)
Nope. They'll have to time out and be reissued.
Actually, I believe there is
)
Actually, I believe there is a way but you would need to install the latest development version to do it so it is absolutely not worth the trouble.
If you check out this thread, you will find a discussion about missing or "ghost" results that were eventually recovered by installing a version of BOINC that could tell the server what it had and therefore the server could work out what was missing and resend it. I imagine this procedure would probably work for you but there are downsides.
1. In order to get the two results that were lost you would be obliged to redownload the large data file that those results were sliced from.
2. Why would you want to do that when you already have a new (and different) data file anyway.
3. The results in question may have already been done by someone else and if a quorum was already formed you'd essentially be wasting your time.
In any case, it's quite painless to let them simply expire on the server. Just kiss 'em goodbye in your mind :).
Cheers,
Gary.
RE: Actually, I believe
)
For Windows systems it works fine with the current production version, BOINC 4.45.
That version isn't available for other operating systems, so for those you'd need one of the later development versions.
Walt
Thanks to all that responded.
)
Thanks to all that responded. I'll take the easy solution and let the issue die peacefully. On another computer I started running E@H and it is maintaining a backlog of 4 jobs ready to run. It's an Intel x64 (not dual core) running at 50%, so it's only processing one job at at time. Having the backlog waiting doesn't seem to be a probelm, but my other 32-bit machine only has one job running and one in process. Is there a setting I need to adjust?