| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Experiencing possible problems running WCG applications on BOINC 7.0.31 using Mountain Lion |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 12
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I have experienced three kernel panics today on my new iMac, which is currently using the OS 10.8.2 and running BOINC 7.0.31, with an AMD Radeon HD 5770 graphics processor. This is the first time I have experienced a kernel panic on this or my previous computer. I know that there are various issues which can cause a kernel panic, including faulty RAM, and I did install RAM recently; however, my tests so far have not indicated any problems with it, nor did the error log from the panic indicate anything associated with the RAM.
I am starting to suspect that it could be WCG projects running on my computer which are causing the problem. This is based on several pieces of evidence: first, the RAM was installed last week, and my computer has run without issue until today, when I received new WCG projects through BOINC. Second, the error logs showed that the processes that it has crashed on are BOINC processes, with the mos recent one starting with wcg. I unfortunately did not save the most recent log, but I do have the projects that were running when it crashed - Help Fight Childhood Cancer 6.40 (one workunit) and FightAids@Home 6.40 (two workunits). I have not confirmed for certain that WCG projects are the culprits, and will need some time in order to do so - I have just restarted from the most recent panic a half-hour ago, and ceased running all WCG projects immediately after viewing the log. Given the rapid frequency of the recent kernel panics, I am going to watch for 24 hours and see if I experience any others. If I do not, I will have to presume that WCG projects are not playing nice with my computer. I don't know if it's incompatibility with 10.8.2, or if it might be associated with the graphics processor (the backtrace highlighted a few extensions that seem to relate to graphics or AMD), or if it's something else. If anyone can provide information as to whether they have experienced these kinds of problems before, or have heard of others experiencing these same problems, please let me know - it could help provide indications to me as to whether this is actually the problem, or if I may be overlooking something. Thank you. |
||
|
|
a_mobile_humanist
Cruncher Joined: May 20, 2011 Post Count: 34 Status: Offline Project Badges: 
|
first, the RAM was installed last week, and my computer has run without issue until today, when I received new WCG projects through BOINC. Second, the error logs showed that the processes that it has crashed on are BOINC processes, with the mos recent one starting with wcg. I unfortunately did not save the most recent log, but I do have the projects that were running when it crashed - Help Fight Childhood Cancer 6.40 (one workunit) and FightAids@Home 6.40 (two workunits). Are you running and having any trouble with GO Fight Against Malaria? Help Fight Childhood Cancer, FightAids, and Go Fight Against Malaria all utilize AutoDock; a possible connection, if these are causing the panics? [Edit 1 times, last edit by a_mobile_humanist at Sep 27, 2012 12:39:31 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I am not currently running any WCG workunits, but I do have two GO Fight Against Malaria workunits that are currently suspended (had not started running before I suspended them). The computer was upgraded very recently, due to hardware failures on my previous computer, and I believe these are the first actual WCG workunits I have received while on this computer (I have been running numerous other projects and not experienced any known problems with any of them thus far), so I don't have much history to look back at, either.
I'm hesitant to start running one, as I have no idea whether I could experience more serious problems if I continue to have kernel panics. However, if there is no other way to test it, I can try forcing BOINC to run those and see what happens. In the meantime, can you provide some information about AutoDock? I am not very familiar with what this is. |
||
|
|
a_mobile_humanist
Cruncher Joined: May 20, 2011 Post Count: 34 Status: Offline Project Badges:
|
AutoDock (http://autodock.scripps.edu) is the software upon which the work units used by the three projects are based. If all other projects are working OK, but it turns out that Malaria does not, in addition to Childhood Cancer and FightAids, then this hypothesis might pan out.
----------------------------------------I mentioned it also because I've been doing some Googling and found some other apparent mentions of AutoDock- But I think you're right and should hold off on trying to run any projects again until the computer runs OK for a while without any additional panics. Should probably also wait for a WCG admin/tech to weigh in as well, as they will know for sure how things are set up to work. Just brainstorming for the moment. [Edit 4 times, last edit by a_mobile_humanist at Sep 27, 2012 1:06:40 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thanks for the information. Yes, there were links to the graphics processor in the backtrace, so that sounds like it could be the cause; it's at the very least an avenue for investigation. However, I will wait and see for the time being, in order to be sure. I'll also look into the other projects I'm linked to and see if any of those use AutoDock, just in case. Thanks.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Update. My computer ran for 16 hours without a fault. The most recent one occurred two hours ago, while I was out. The error pointed to another BOINC application, NFS@Home. I'm trying to ascertain whether this project uses Autodock; there doesn't seem to be much information out there about which projects do and do not use that. There were no other backtraced extensions in the log this time, so nothing pointing to the graphics processor.
My suspicions are even stronger that Autodock is the culprit for the WCG units. By some fortune I found the log of the third kernel panic yesterday - apparently I hadn't closed it, it just got hidden behind other windows - and the process that it faulted on was wcg_faah_autodoc. The faults have always occurred after recent acquisition and activation of BOINC WUs from a project that hasn't run yet on this OS. And it's only a limited number of projects - I've been successfully running WUs from eon2, rosetta@home, climateprediction.net, SETI@Home, SZTAKI Desktop Grid, yoyo@home, Einstein@home, sudoku, malariacontrol.net, MilkyWay@Home, Poem@Home, and Primegrid, all of which I have run subsequent to updating to 10.8.2. I have not erased the WUs from the offending projects, but I have suspended them all and am currently blocking new tasks from being received from the projects. I am hoping that there will be either an update to Autodock or to OS X that resolves these issues. If anyone can find a list of all the BOINC projects that use Autodock, I would be grateful. So far my searches have not turned up much - I've seen mention of the WCG projects running Autodock, and one other project that I am not attached to. I am still not sure about NFS, nor about any of the other projects I run. If I can stop them from coming in, it would be preferable to waiting for kernel panics. Any help on this is greatly appreciated. Thanks. |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges: 
|
----------------------------------------  - AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Have read already. I am already running 7.0.31.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The only 2 sciences at WCG that use AutoDock are FA@H and HFCC. The plan is to upgrade the AutoDock version for FA@H. That said GFAM, now running on VINA, will be getting a second reprocessing pass on the new version of AutoDock. Not heard for a long time where we are on that, but last was that it was not sure which of the 2 [FA@H or GFAM] would have a first try on this upgraded docking engine.
If it's AutoDock or ML, the reports are so far limited to this thread. Identification of the science app used: Usually, at least at WCG, the application name includes a reference which one is used, but not it's version. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thanks for the info. I've been searching online for further information, without much success, but I've seen numerous reports of frequent kernel panics on machines running 10.8.2. I am currently considering two possibilities:
One, that 10.8.2 has compatibility issues with certain BOINC projects, which have not been addressed due to the recency of its release and possibly because there simply aren't many people running these workunits on this version of the OS and thus no reports. The faults with NFS appear to be unrelated to the WCG faults, as NFS does not run Autodock, but I don't think it unlikely that there could be compatibility issues with more than one process. Two, that there is some legacy information interfering with the running of these applications on 10.8.2, which is possible due to the fact that this computer's data came from a Time Machine restore from another computer with an earlier OS. However, I'm not sure what this could entail, as there is nothing clearly linked to these faults other than these two BOINC projects. I am continuing to explore and will provide updates as they come. Currently have four hours of uptime on the machine without incident, and keeping an eye on any new WUs that come in to BOINC. |
||
|
|
|