Make sure heap dump file survives the pod restart

[Tristan1900]

Description

We are experiencing OOM error in segmentstore. When that happens the heap dump file is generated and stored locally in that pod and the pod is restarted instantly, so we lost the heap dump file for debugging.

Importance

Should have

Location

Segmentstore and bookkeeper code

Suggestions for an improvement

Use an emptyDir to store the heap dump file and specify the corresponding heap dump directory to that emptyDir.