Follow-up from "Stuck loop"
The following discussion from !3409 (merged) should be addressed:
-
@DavidVorick started a discussion: (+3 comments) Between the health loop and the stuck loop, do we have two different threads running repairs?
I'm not sure that the repair loop was designed to be safe to have multiple threads unloading chunks at the same time, all of the chunks from the stuck loop and repair loop should be going into the same heap and then a different thread should be working through the heap.
TODO
-
Review Health, Repair, and Stuck loop flow. Document and optimize if possible.