I updated the VM to latest changes and rebooted it .. so far it's looking as if things seem to be working nicely now ... let's hope it stays that way. If not I'll request more RAM for our VM.
Chris Am 19.03.19, 13:09 schrieb "Christofer Dutz" <[email protected]>: Hi all, I think I just found out why we have so many cancelled builds … I was puzzled about why the jobs are sort of randomly cancelled and especially every time more than one build running in parallel. When looking at the “dmesg” output on our agent I noticed several entries like this: [15967108.392372] Out of memory: Kill process 30822 (java) score 294 or sacrifice child [15967108.396284] Killed process 4000 (javadoc) total-vm:3349304kB, anon-rss:58084kB, file-rss:0kB When executing the “top” command I can see about 4 java processes … will wait till the current build is done and kill the rest as I assume they are zombies. Will stay on the topic and try to solve this problem as memory is really not a PLC4X problem (We’re running with the java defaults which should not be much) Chris
