The donor shard is first committing the migration on the config server and then waiting for orphans to be deleted.
The _configsvrMoveRange command calls into Balancer::moveRange that retrieves the current shard from config.chunks and returns early in case the range is already present on the recipient shard.
Given the above mentioned conditions, in case the config server command is retried it may happen that it succeeds before the donor finishes waiting for range deletions.
- is caused by
-
SERVER-65831 BalancerCommandsScheduler::stop() should not delay the execution of the Config Server step up routine
- Closed
- is related to
-
SERVER-46669 moveChunk may succeed but not respect waitForDelete=true if replica set shard primary steps down
- Closed
- related to
-
SERVER-70852 Recommit SERVER-66716 on v6.0
- Closed