SnapMirror transfer fails for large volumes with 'CSM: An operation did not complete within the specified timeout window'
Applies to
- ONTAP 9
- SnapMirror
- SVM DR
- SMTape (SnapMirror to Tape)
- Memory exhaustion
Issue
- SVM DR transfer fails for a volume with the error:
Last Transfer Error: Transfer for volume "Volume_Name" failed. Reason: Transfer failed. (Error marshaling replication operation data on Cluster Id: 2f6e64aa-95df-12e6-8bce-00a022a9226d, Node Id: 44da037c-94de-15e6-a8c8-cff55d91aa61 (idl out of memory)).
- SnapMirror Transfers fails for large volumes and reports a timeout in the
SnapMirror audit logs
Fri Mar 18 09:05:04 CEST 2020 ManualUpdate[Oct 9 09:00:54]:9863154b-77e1-11e8-8090-00a09864b794 Operation-Uuid=2cab1a60-09fd-11eb-b869-00a09867f2a6 Group=none Operation-Cookie=0 action=Defer source=svm1:vol1 destination=svm2:vol2 status=Failure message=Transfer failed.(Replication operation request failed from Cluster Id: "Source Cluster Id", Node Id: "Source NodeId" to Cluster Id: "Destination ClusterId", Node Id: "Destination NodeId" due to a network error(CSM: An operation did not complete within the specified timeout window.))
- On
Sktrace logs
low FreeBSD memory errors will be logged
2021-03-18T09:00:02Z 39696815208071979 [3:0] GFC_FC: gfc_vm_lowmem_start_flowcontrol: started global flow control due to low FreeBSD memory
- Sometimes snapmirror out-of-memory errors are reported in
EMS logs :
Fri Mar 18 06:05:10 +0100 [n1: CsmMpAgentThread: repl.out.of.memory:notice]: SnapMirror replication transfer encountered an out-of-memory error.
- SMTape will show this failure on the
EMS.log
:
Wed Mar 20 12:02:50 +0100 [Node01-n01: ib_bkp_main: smtape.bkp.fail:error]: SMTape backup session 6000012 from /SVM-Name/Volume_name to NDMP_LOCAL failed with error failed to marshal replication operation (idl_error = 5) 22:92.