[CONTROLLER-1158] Clustering:Fix in-mem log cleanup for an Inactive follower Created: 13/Feb/15 Updated: 21/Feb/15 Resolved: 21/Feb/15 |
|
| Status: | Resolved |
| Project: | controller |
| Component/s: | mdsal |
| Affects Version/s: | Post-Helium |
| Fix Version/s: | None |
| Type: | Bug | ||
| Reporter: | Kamal Rameshan | Assignee: | Kamal Rameshan |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Environment: |
Operating System: All |
||
| External issue ID: | 2715 |
| Description |
|
Currently the in-mem log entries will get removed only if those entries are replicated to ALL the nodes. This is true for both real and fake snapshots. If one of the followers is down, the log would not get cleared causing the system to become unstable. Also during rolling upgrades, if we take a follower down, we do want the rest of the system to function smoothtly. We might want to come with a leader tolerance factor , beyond which the leader would start cleaning the log even if its not replicated to all. |
| Comments |
| Comment by Kamal Rameshan [ 16/Feb/15 ] |
| Comment by Kamal Rameshan [ 16/Feb/15 ] |
|
Currently the simplest fix for this would be to go for the clearing if the memory goes beyond the threshold. Same as the check for Snapshot in RaftActor |