Hi all, I am not a Java expert, but yesterday we had a bloomreach site go down due to the number of Java threads just went sky rocketing…and we had out of memory issues on the containers.
The rest looked okay…
Can anyone shed some light on jvm threads… what might cause this…?
I just restarted the nodes and all was fine…
J
Hi all, any one got any thoughts on this…?
We run bloomreach in a docker container and use the defaults.
ENV JAVA_MINHEAP 256m
ENV JAVA_MAXHEAP 512m
ENV TOMCAT_MAXTHREADS 200
We also allow 2gb Ram to the container…
Is it worth tweaking the variables to increase the memory allocation and use more of the container allocated memory?
That’s hard to say… the defaults normally work well.
Can you figure out what has caused the OOM by inspecting the memory dump (.hprof file) that has been generated when the OOM occurred? At least if you have specified -XX:+HeapDumpOnOutOfMemoryError on the JVM.
Jeroen
Hi yes that option is specified, however due to other issues, folder mapping is not exposing the temp folder, so that memory dump is lost 
