At random times, fsm_workloads/snapshot_read_at_cluster_time_crud_operations.js searches currentOp for getMore operations and calls killOp on them, which causes them to fail with error code 11601 Interrupted. However, the test expects getMore to fail only with error code 91 ShutdownInProgress.

I think the test should expect getMore to fail with error code 11601 Interrupted. The test flaw leads to BFs rarely because the getMore doesn't last long, so when the test searches for getMores to kill it rarely finds any. When the test passes it doesn't log "Going to kill op" at all.