E@H Linux problems
23 Jul 2005 13:01:43 UTC
Topic 189594
(moderation:
I just started doing E@H yesterday and I've had to detach the E@H project from my linux box. It seemed whenever it switched to E@H, the process would just sleep with a message, "Deferring communications for XX hours", but didn't ever seem to do any actual work.
Seti, Climate, and Protein all seem to work fine. Einstein was the only one that was problematic.
I ended up detaching the E@H project and now my Linux box has gone back to being busy all the time with Boinc projects. I'd like to have it work on E@H, but if it's just going to sleep, then it's not worth it.
As for the version, I believe it's einstein_4.80_i686-pc-linux-gnu
Pete
[edited content: Had an E@H windows problem as well, but found the solution]
E@H Linux problems
)
Could be a number of things. Checked your firewall settings? Looking thru the logs (usually at /var/log/) for rejected/denied IP packets may be a good idea.
Or maybe the Einstein server was inaccessible for some other reason just then. Like f.i. saturated LAN.
Or maybe your options setup with einstein.phys.uwm.edu are cooked? The "Use network only between the hours ..." option comes to mind.
Anyway, IMHO it's better to concentrate on 1 or 2 projects than to spread one's computing resources thinly.
Greetings, Ragnar Schroder
RE: I just started doing
)
Pete, according to my server logs, the last time that your Linux box contacted the E@H scheduler was July 23rd. At that time it had a full quota of 8 workunits. My logs show no contact since that time. Database shows that one result failed with an error (0 CPU time used) and the other results timed out (no response).
Also, you appear to be running boinc client version 4.19. I suggest you consider upgrading to a more recent Linux version of BOINC (both client and client manager).
Bruce
Director, Einstein@Home