I'm using Ramdisk to run Bionc, every once in a while my computer freezes/crashes and I loose my ramdisk or it doesn't copy the disk image fully before the crash. There should be a way to request a re-download the lost work units from the server/s.

---

I see your point, but just wanted to point out that the project lives on. I mean the WU that you (due to your crash) have become unable to report back will eventually reach it's deadline, and will then be sent to someone else that's requesting work. This is normal, and the project handles it just fine. I mean, no harm is done, other than you've lost some productive crunch time. And THAT will be the case regardless.

Since it is possible (indeed, it apparently just happened this weekend) that the WU itself is part of a problem in processing, if you continually download the same one, you'd be stuck with the same problem longer than if you just get a new piece of work assigned. The "new" work, would have the problem resolved, and therefore get you crunching productively sooner than if you retry the existing WU you had.

I get the sense that you are hoping to sort of clean up the "mess" that resulted from your crash, but there's really no need.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

dont use ramdisk... thats an old, unnecessary thing in this day and age.

---

I just experienced a bad WU. The message log told me to reset the project to clear it. I should of suspended it and let my manager complete the other WUs, but I did not, so I lost three WUs.
Everybody is saying..."big deal..the project goes on...the WUs will just time-out and go to someone else". I think this is inefficient.. now those three WUs are just sitting there, waiting until they expire (10 days from now) before they get resent to someone else.
If there was a link to redownload the WU, maybe on the WU ID page or on my computer page or something..then I would be able to complete those two good WUs I lost..and whomever is chosen to verify the WU will not have to wait, in this case 10 additional days, for their pending credits to be given them....and the project gets the result 10 days faster.

Rosetta doesn't make you wait for credit, nor burn a second run of the identical work just to verify your results. So these tasks aren't holding up anything.

Bottom line is that every single scheduler request would have to parse through your report and then read through the database for ALL of your existing WUs to see if there are any that were not reported as "in progress" on your machine. It is a fairly expensive operation to do for 100,000 machines. So, while you are certainly correct that there is some inefficiency, and lack of beauty in not assigning that work back to you, it is the lesser of two evils.

---

Rosetta Moderator: Mod.Sense

Ah. :) Still new and am used to running Seti, so I did not realize Rosetta did not verify results. Thanks. :)