influxdata / influxdata/influxdb
Add flag to DefaultPlanner to bypass the wait for shards to become cold.
Open
1.x
area/tsm
team/edge
- Dominant language
- Rust
- Stars
- 31.7k
- Forks
- 3.7k
- Avg merge
- 13h 37m
- Merged PRs (30d)
- 8
Description
__Proposal:__
The DefaultPlanner object should store an internal flag complementary to forceFull which causes the time check against `compactFullWriteColdDuration` to be skipped. This would allow compactions after an Enterprise anti-entropy repair to be scheduled immediately, speeding the repair process.
__Current behavior:__
AE repairs cause shards to become hot, and they then have to wait until the cold duration in order for the further repairs to be processed.
__Desired behavior:__
AE repairs should not have to wait for the cold duration before compaction, so the process can continue.
Contributor guide
Assessment
This issue has not been assessed yet.