Memory tuning for big projects

[shusterboris]

Dear colleagues!
I encountered a memory shortage. I have 12 large projects. Indexing takes 8 hours or more. The size of the used heap slowly grows, and eventually the process is killed. And most often the "fattest" process is killed, i.e. the root one. After that, the container has to be rebooted, indexing starts again - and so on in a circle. The server has 8 cores. OpenGrok starts 16 threads. It seems that the problem can be solved by adjusting parallelism. But how to set it for the use case when indexing starts at service startup. Adding the following setting to the configuration file does not have any effect.

4

I tried changing the container settings, but I couldn't get it to work reliably.
docker create --restart unless-stopped -p 8080:8080 --env SYNC_PERIOD_MINUTES='0' -m 30g --cpus="7.5" --env JAVA_OPTS="-Xms2g -Xmx6g -server" --env CATALINA_OPTS="-Xms7g -Xmx36g" ...
If you run indexing of projects one by one, it will work, but how to survive the general reindexing at launch
I would be very grateful for advice

[shusterboris]

I partially solved the problem with memory usage. Well, not solved... bypassed. I reduced the maximum heap size for JAVA_OPTS and CATALINA_OPTS to 4 GB, removed the launch of reindexing after restarting the container (by changing docker/start.py) and launch reindexing of individual projects one by one on a schedule