Page 1 of 1
Disk limit
Posted: Sat Jul 09, 2011 12:56 pm
by Stef42
I'm getting errors with max disk usage exceeded, only 10MB and the WU is 10,50MB. I think the WU parameters for disk usage should be raised a bit?!
Re: Disk limit
Posted: Sat Jul 09, 2011 3:33 pm
by GaryWilson
Same here, but looks like its only my Windows machine. I was completing a bunch of 3.06 tasks, but then it just started happening. I complete an occassional task though. So not all tasks are hitting the projects internal 10 mb file limit.
Re: Disk limit
Posted: Sat Jul 09, 2011 3:36 pm
by GaryWilson
Actually I take that back. It's happening on my Linux machines as well. And it's only recently, so the first batch of 3.06 tasks went fine, its just the latest ones that are doing this. My Mac is still running the 3.05 client, so looks like the Mac client isn't updated to 3.06 yet.
Re: Disk limit
Posted: Sat Jul 09, 2011 7:44 pm
by Stef42
The large units that I get crash with the disk limit error, the small one's with runtime of around 2000sec. finish just fine.
Re: Disk limit
Posted: Sun Jul 10, 2011 6:46 pm
by NATE1
yep, same here,
Seems the best thing to do is on Friday make sure you have enought work units, if you can, to last till Tuesday, run bonic in the background, and not even give it a 2nd look. I running over 40 projects and it's always the same, every weekend. boom one or more will go west.
Re: Disk limit
Posted: Sun Jul 10, 2011 10:24 pm
by Conan
All work units failing with "Maximum Disk Usage Error". Linux and Windows.
Have set to NNW.
Conan
Re: Disk limit
Posted: Mon Jul 11, 2011 1:14 am
by chill
Sorry about this problem. I have increased the disk limit temporarily. I will decrease the disk usage in a future version of the client.
Re: Disk limit
Posted: Mon Jul 11, 2011 6:23 pm
by Stef42
Errors fixed but no work available! What's going on?
Re: Disk limit
Posted: Mon Jul 11, 2011 6:41 pm
by chill
A new algorithm in the server side code that compares clusters of atoms is taking up all of our cpu time. It is taking too much time to process and compare the incoming workunits and thus no new work is being created. I started up a new simulation that shouldn't have this problem.