Posts by [B^S] ThatGuy

1) Message boards : Number crunching : Serverproblems ? (Message 10829)
Posted 16 Feb 2006 by Profile [B^S] ThatGuy
Post:
Thanks for the info!

I'm having problems downloading to one of the computers that I have running Rosetta, but it is only certain WUs that have a problem. I've had many newer WUs come through fine.

I did a manual "Retry Now" on each of them, and I noticed a common denominator to the files that will not transfer - The actual files are larger (most of them 10-20x) than the "expected" size. My hypothesis is that the file gets downloaded, then a check happens to make sure that the file is "good", but the size / checksum doesn't match expected values, so it fails the transfer. Not much of a reach, I know.

So... do I abort the transfers, or should I turn off the validation? Would that actually help? More importantly, how in the world did the expected sizes become different than the real sizes?

EDIT: There is nothing in the message log that really indicates what is going on - just "Temporarily failed transfer".


This happens sometimes. It is not really an error in the file size, (though you are correct that it looks like one) it is a problem on the server side, usually a dropped connection. Usually, in time, these will sort out by themselves, but you can stop and start BOINC and that will "sometimes" get them going. In any case if left alone they will eventually download by themselves.

2) Message boards : Number crunching : Serverproblems ? (Message 10817)
Posted 16 Feb 2006 by Profile [B^S] ThatGuy
Post:
I'm having problems downloading to one of the computers that I have running Rosetta, but it is only certain WUs that have a problem. I've had many newer WUs come through fine.

I did a manual "Retry Now" on each of them, and I noticed a common denominator to the files that will not transfer - The actual files are larger (most of them 10-20x) than the "expected" size. My hypothesis is that the file gets downloaded, then a check happens to make sure that the file is "good", but the size / checksum doesn't match expected values, so it fails the transfer. Not much of a reach, I know.

So... do I abort the transfers, or should I turn off the validation? Would that actually help? More importantly, how in the world did the expected sizes become different than the real sizes?

EDIT: There is nothing in the message log that really indicates what is going on - just "Temporarily failed transfer".
3) Message boards : Number crunching : Report stuck & aborted WU here please (Message 10650)
Posted 11 Feb 2006 by Profile [B^S] ThatGuy
Post:
I got home to find this one stuck at 1% after over 20 hours:

http://boinc.bakerlab.org/rosetta/workunit.php?wuid=8234872

PRODUCTION_ABINITO_CENTROID_PACKING_1nspA_301_1457_0







©2024 University of Washington
https://www.bakerlab.org