Memory Leak

[ShrimpCryptid]

The SHO server has a set amount of RAM available to it. However, the amount of RAM used by the program seems to slowly creep upwards over the lifetime of the server, until it eventually hits the upper limit and the hosting service terminates it.

The cause of the memory leak should be profiled (dump Java memory log, etc.) and fixed. Another solution would be restarting the server on a set interval.

[ShrimpCryptid]

Update: I've added some crash logging that will hopefully activate the next time the server runs out of memory! Hopefully, I can get some useful information out of the heap dump.

[ShrimpCryptid]

Thought: SHO does a LOT of logging. Is it possible that the log/terminal session never clears old output, and that's what's taking up system memory?

Test:

• Test deleting/minimizing most if not all logs. Does the rate of memory usage slow?
• GitHub Action to periodically re-launch the server (will this cause user interruptions?)

[ShrimpCryptid]

Okay, looks like the program never did do a heap dump even when it hit the memory limit. I've recorded the rate of memory usage (it seems to increase by roughly 23 MB/day) and I'll see if the rate decreases once I remove most of the logging. There are a few other exceptions getting thrown due to the Jetty API updating that I also need to address.

[ShrimpCryptid]

Okay, so post-fix it looks like the rate of memory increase has dropped to 3 MB/day. I need to look into Java logging libraries but for now logging does seem to be the main culprit.