| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Mapping Cancer Markers Forum
Thread: Neverending MCM tasks |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 86
|
|
| Author |
|
|
Steve Breibart
Cruncher Joined: May 11, 2007 Post Count: 2 Status: Offline Project Badges: 
|
Running on Windows 10 using the BOINC manager. MCM tasks just appear to keep going having reached 100%. A temporary fix is to reboot Windows but I can't keep doing that all the time.
|
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges: 
|
Running on Windows 10 using the BOINC manager. MCM tasks just appear to keep going having reached 100%. A temporary fix is to reboot Windows but I can't keep doing that all the time. Long standing issue with MCM, this is going on (randomly for me) for years. well before the IBM-2-Krembil move. And you do not need to restart Windows (at least I didn't have to), exiting out of BOINC (including stopping tasks) and restarting BOINC should do it, so most of those WUs will result in a computation error anyway, so I got used to just aborting those (where I can see them, on remote hosts they might unfortunately run until they time out past their deadline). This apparently was never an issue for the WCG/project folks of old so they never looked into it/fixed. At least for me, it is usually only one or two out of 6 or 8 tasks simultaneously running, and those other WUs ran just fine around those "stalled at the finish line" ones... Ralf |
||
|
|
ca05065
Senior Cruncher Joined: Dec 4, 2007 Post Count: 328 Status: Offline Project Badges: 
|
Another solution worth trying for a stuck work unit is:
set leave application in memory off suspend the work unit wait a few seconds so it is removed from memory resume the work unit which should restart from the last checkpoint taken set leave application in memory on |
||
|
|
Nick Batos
Cruncher Joined: Apr 5, 2020 Post Count: 7 Status: Offline Project Badges: 
|
Well, I am having the issues with MCM units as well. They 'crunch' for less than 2 minutes, then restart- that is the percent complete goes back to 0%. This happens in a never ending loop. This is occurring on both my iMac Pro (using Intel Xeon W CPU, RAM is ECC), and on my new iMac Pro laptop (using Apple M1 Pro CPU). Both running macOS Monterey 12.5.1. All other BOINC projects work, including other World Community Grid projects). Given all of that, my experience is that there is an issue with MCM
|
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges:
|
Well, I am having the issues with MCM units as well. They 'crunch' for less than 2 minutes, then restart- that is the percent complete goes back to 0%. This happens in a never ending loop. This is occurring on both my iMac Pro (using Intel Xeon W CPU, RAM is ECC), and on my new iMac Pro laptop (using Apple M1 Pro CPU). Both running macOS Monterey 12.5.1. All other BOINC projects work, including other World Community Grid projects). Given all of that, my experience is that there is an issue with MCM Well, have over the years never seen those exact symptoms, neither on Windows, Linux nor macOS.So I have occasionally that the WU gets stuck (well) below 100%, most of the time with no time shown under "Remaining". It seems that all happens when the WU hits a "sub-job" (for lack of a better word right now) and is retrying the same part over and over again. So when I look at the CPU used on that task, it usually shows (almost) zero. As mentioned, I reported this in the past several times, and while otherwise the WCG techs have been pretty responsive, I never got a direct reply like "we are going to look into this", just an indirect note once that "this doesn't happen often enough in the grand scheme of things" that it didn't see worth to look into it. So I did usually cut my losses and just about those WUs, it usually doesn't make a dent as far as I am concerned, just a bit aggravating at times... As for now, I would much rather see that Krembil focuses on getting all the other pending issues that prevent from WCG overall running at full capacity again before "hopefully" looking into this issue at a later point... Ralf [Edit 1 times, last edit by TPCBF at Sep 1, 2022 2:32:33 PM] |
||
|
|
MyrCu
Cruncher Joined: Apr 9, 2020 Post Count: 49 Status: Offline Project Badges:
|
Today 6 MCM WU had been "Server aborted" (and 1 ARP).
|
||
|
|
Felix Kaeufer
Cruncher Joined: Feb 3, 2012 Post Count: 29 Status: Offline Project Badges: 
|
Just a short update: The issue on macOS systems is still around.
|
||
|
|
mikaelwigander
Cruncher Joined: Apr 2, 2020 Post Count: 3 Status: Offline Project Badges:
|
I have two Mac computers running BOINC and the one running OS 12.5.1 (21G83) is not working properly, the same issues as stated but for my other older one running OS 10.11.6 MCM works fine.
I had to download an older BOINC distribution in order for it to be installed on my old Mac but the version of them both states 7.20.2 |
||
|
|
Jorlin
Advanced Cruncher Deutschland Joined: Jan 22, 2020 Post Count: 90 Status: Offline Project Badges:
|
Once had the problem of WUs getting stuck at 100%.
----------------------------------------In my case the cause was an anti malware program. After excluding the BOINC folder from scans everything was fine again.  |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
What I'm seeing lately are MCM WUs that get stuck at around 27% indicating that 2 days have been crunched and aprox 6 days remaining. It doesn't happen very often (about once a week) and I just abort the WU.
----------------------------------------Thanks, CJSL |
||
|
|
|