In the migration_recovers_unfinished_migrations.js test there is a stepUp and then a moveChunk is called in a parallel shell.
If the source shard is still executing the step up process, the moveChunk command will fail with a NotWritablePrimary and won't be retried due to SERVER-74177, therefore the test will fail.
The purpose of this ticket is to wait for the primary to reach a steady state before resuming the moveChunk command.
- related to
-
SERVER-74177 balancer_commands_scheduler wraps retryable error into OperationFailed error, making them not eligible for retry
-
- Closed
-