Message boards : Number crunching : WU Problem Linux
Author | Message |
---|---|
Chilean Send message Joined: 16 Oct 05 Posts: 711 Credit: 26,694,507 RAC: 0 |
|
KSMarksPsych Send message Joined: 15 Oct 05 Posts: 199 Credit: 22,337 RAC: 0 |
What version of BOINC? Even the latest alpha is woefully behind the times as compared to Windows and Mac. :( I think I just ran into the exact same issue last night with CPDN beta. Kathryn :o) The BOINC FAQ Service The Unofficial BOINC Wiki The Trac System More BOINC information than you can shake a stick of RAM at. |
zombie67 [MM] Send message Joined: 11 Feb 06 Posts: 316 Credit: 6,621,003 RAC: 0 |
On my LInux box... once ina while i get workunits that are supposedly "running" but really arent. This happened to me all the time back when my machines ran linux. It required constant monitoring to make sure they weren't sitting idle. That, combined with lower credit claims, pushed me back to windows. This was about a year ago. Reno, NV Team: SETI.USA |
FluffyChicken Send message Joined: 1 Nov 05 Posts: 1260 Credit: 369,635 RAC: 0 |
While the credit claims is not a problem in Rosetta and many other projects now. I've had my fair share of problems under linux. I've use 5.4.x and 5.8.x both in Ubuntu 7.04 (edubuntu actually) I have played around before that but cannot remember exact problems. Now things to look out for are. - SpeedStep/Cool&Quiet techniques kicking in, Often mines sat at 600MHz (Pentium-M) when the computer is idle and boinc is running. I have to override it and force it to full speed. - Memory, which maybe what's happening to you. Check your setting for memory percentage to use. If Rosetta starts to go over that amount it will just stop, drop to zero processor use and well do as you say. Happens in Windows too btw. Check in BoincManager and the message to see what s happening. Other things like it just quiting for no reason, but that's not happened with 5.8.x version. P.S. If you unhide your computers, we can look at the unit's themselves and you computer stats. Team mauisun.org |
Chilean Send message Joined: 16 Oct 05 Posts: 711 Credit: 26,694,507 RAC: 0 |
While the credit claims is not a problem in Rosetta and many other projects now. But of course. Now that I remmeber... this happened on a windows machine. Only that i saw a message saying "waiting for memory". SO I bumped the memory form 50% to 80%. Thanks for the solutions!! |
DJStarfox Send message Joined: 19 Jul 07 Posts: 145 Credit: 1,250,162 RAC: 0 |
On my LInux box... once ina while i get workunits that are supposedly "running" but really arent. I have the same problem, but it's not due to memory. The program just sits in a QUIT_PENDING state and will not exit. Then when BOINC goes to resume the task, it just hangs. I have to kill the pid, and then BOINC will restart it properly. This has been a problem since day 1 and it does not matter whether leave-apps-in-memory setting is yes or no. |
mray Send message Joined: 3 Dec 06 Posts: 4 Credit: 33,856,665 RAC: 1,591 |
I was having this problem on a Linux Mepis box. I updated to the latest version of BOINC and the problem seems to be reduced or perhaps even fixed. I haven't caught it sleeping on the job in a while at least. I think I'll check my memory allocation though, I have seen that "Waiting for memory" error even though the machines should have plenty of meemory for the task. |
DJStarfox Send message Joined: 19 Jul 07 Posts: 145 Credit: 1,250,162 RAC: 0 |
I was having this problem on a Linux Mepis box. I updated to the latest version of BOINC and the problem seems to be reduced or perhaps even fixed. I haven't caught it sleeping on the job in a while at least. I think I'll check my memory allocation though, I have seen that "Waiting for memory" error even though the machines should have plenty of meemory for the task. I'm running BOINC version 5.8.16 for i686. I haven't upgraded to 5.10.* because they are still working out serious bugs last time I checked. |
computerguy09 Send message Joined: 17 Sep 05 Posts: 2 Credit: 5,869,369 RAC: 34 |
I was having this problem on a Linux Mepis box. I updated to the latest version of BOINC and the problem seems to be reduced or perhaps even fixed. I haven't caught it sleeping on the job in a while at least. I think I'll check my memory allocation though, I have seen that "Waiting for memory" error even though the machines should have plenty of meemory for the task. I saw this problem on my Ubuntu box (dual core AMD) while running 5.8.x, and early versions of 5.10.x. I've been running 5.10.8 for over 2 weeks now with no "hanging" processes. Used to have to restart BOINC every day or so on this (and other) boxes, but haven't seen it since going to 5.10.8 of BOINC. Mark |
DJStarfox Send message Joined: 19 Jul 07 Posts: 145 Credit: 1,250,162 RAC: 0 |
Hmm...well, let me download 5.10.8 or higher and check it out for myself. It's still considered beta, so I'd rather finish my current climate model first. Seems fishy that only project Rosetta would not exit cleanly; all other projects perform this fine (Climate, SETI, Einstein). |
Message boards :
Number crunching :
WU Problem Linux
©2025 University of Washington
https://www.bakerlab.org