| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: OPN1 WU memory leaks? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 41
|
|
| Author |
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 2220 Status: Offline Project Badges: 
|
Thank you knreed.
|
||
|
|
paulch2
Cruncher Joined: Aug 6, 2020 Post Count: 25 Status: Offline Project Badges: 
|
Odd.. I keep seeing jobs error because disk limit exceeded 200MB
No idea where that limit is. Disk limit for total jobs is 777GB so not related to any obvious web site settings |
||
|
|
paulch2
Cruncher Joined: Aug 6, 2020 Post Count: 25 Status: Offline Project Badges:
|
Ignore that. Must be a local problem. I have another machine with identical hardware, Same OS and Client version, and running similar batches of OPN1 tasks which is fine. Time for a reboot...
|
||
|
|
flynryan
Senior Cruncher United States Joined: Aug 15, 2006 Post Count: 235 Status: Offline Project Badges: 
|
Running 32 threads of OPN, seeing usage between 266 MB and 2.6 GB per work unit! I have 32 gb on each machine so haven't run out of memory yet but its definitely massive usage.
|
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
Odd.. I keep seeing jobs error because disk limit exceeded 200MB No idea where that limit is. Disk limit for total jobs is 777GB so not related to any obvious web site settings The 200MB disk limit is set as a constraint on the job assigned to you. If that limit is exceeded, BOINC should kill the job and notify the server which it appears to be doing. |
||
|
|
KerSamson
Master Cruncher Switzerland Joined: Jan 29, 2007 Post Count: 1679 Status: Offline Project Badges:
|
I've just read the posts in this thread after having commented on knreed's thread (see below)..
------------------------------------------- Currently I do not observe any memory troubles on my Linux machines, even not on Windows 7; but it "killed" today two Raspberry Pi 3. RAM 97%, swap space 100% (with 512 MB swap space). --- Cheers, Yves ---------------------------------------- [Edit 1 times, last edit by KerSamson at Oct 5, 2021 8:01:45 PM] |
||
|
|
spRocket
Senior Cruncher Joined: Mar 25, 2020 Post Count: 277 Status: Offline Project Badges:
|
All of my PC-class machines have sufficient RAM to handle the tasks they get, but my Raspberry Pi systems are gagging. I'm seeing one or more "Waiting for memory" tasks while these things percolate through the system, but everything seems to eventually get processed. I'm using zram for swap to avoid killing my SD cards and still provide a cushion when RAM gets low.
|
||
|
|
TonyEllis
Senior Cruncher Australia Joined: Jul 9, 2008 Post Count: 261 Status: Offline Project Badges: 
|
The swap on my Pis (mixture 3B+ 3A+ and Zero) are 2G+ and they seem to be running OK. No WUs failures so far... All boot from SD (just boot partition) and the rest of the OS etc on USB connected SATA SSD or HDD. The rest of the SD is used to contain the daily backup, in addition to backups to NAS.
----------------------------------------Other systems similarly have very large swap areas and are surviving...
Run Time Stats https://grassmere-productions.no-ip.biz/
|
||
|
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 988 Status: Offline Project Badges: 
|
knreed has now confirmed that there is a leak of about 14MB for each job within a work unit; it's a good job the tasks with 200+ tasks were scarce!
Apparently there will be a temporary work-around (smaller work units!) and a proper fix later -- it will be interesting to see whether the leak is in OPN1 itself or in its interactions with the BOINC support logic (checkpoints?)! Cheers - Al. |
||
|
|
|