importccl: --experimental-direct-ingestion creates underreplicated ranges #37014
Labels
A-disaster-recovery
C-bug
Code not up to spec/doc, specs & docs deemed correct. Solution expected to change code/behavior.
S-3
Medium-low impact: incurs increased costs for some users (incl lower avail, recoverable bad data)
I ran
on a 3-node cluster on AWS. After this had completed, I ended up with around 6k ranges, almost all of which were underreplicated; there were slightly fewer than 6k replicas on 1 node and around 100 replicas on each of the other two nodes.
I haven't been able to consistently reproduce this.
The text was updated successfully, but these errors were encountered: