Nodes with shouldhaveshards=false should not participate in repartition joins
Open
- Dominant language
- C
- Stars
- 12.8k
- Forks
- 794
- Avg merge
- 2d 14h
- Merged PRs (30d)
- 31
Description
Right now when you do a repartition join even the coordinator is included. This is quite unexpected and probably not intended. The coordinator might have very different resources from the workers (disk space/cpu/max_connections).
Contributor guide
Assessment
This issue has not been assessed yet.