The in memory state of the defragmentation policy isn't being cleared on step down. We are stopping the balancer threads, which prevents defragmentation from continuing after the node has stepped down, but leaving the in memory state intact can cause defragmentation to resume on re-step-up of an old primary when it should not be resumed (for example, when defragmentation has completed with a different primary). As part of the step down of the balancer, we should clear the _defragmentationStates map in the balancer defragmentation policy.
- Votes:
-
0 Vote for this issue
- Watchers:
-
3 Start watching this issue
- Created:
- Updated:
- Resolved: