MTUの不一致が原因で、ktls.badAuthのCSMエラーによりSnapMirror処理が失敗し続ける
環境
- ONTAP 9
- SnapMirror
問題
- SnapMirrorの動作が次のCSMエラーで失敗します。
CSM: An operation failed due to an ONC RPC failure.
CSM: Operation referred to a non-existent session.
CSM: Connection aborted.
CSM: Request was canceled because of a connection failure.
SNAPMIRROR-AUDIT-LOG-TXT.GZ
宛先でCSMエラーが繰り返し発生
status=Failure message=Transfer failed.(Replication operation request failed from Cluster Id: <SOURCE_CLUSTER_ID>, Node Id: <SOURCE_NODE_ID> to Cluster Id: <DESTINATION_CLUSTER_ID>, Node Id: <DESTINATION_NODE_ID> due to a network error(CSM: Request was canceled because of a connection failure.))
EMS-LOG-FILE.GZ
デスティネーションではcsm.sessionFailed
とktls.badAuth
両方または両方が同時に表示されます。
[<NODE>: CSM_BTLS: ktls.badAuth:notice]: ONTAP received a TLS message with incorrect authentication at local address <DESTINATION_INTERCLUSTER_LIF> local port 24278 from remote address <SOURCE_INTERCLUSTER_LIF> remote port 11105 via IPspace -1.
[<NODE>: kernel: csm.sessionFailed:debug]: Cluster interconnect session (req=<NODE>:dblade, rsp=<NODE>:dblade, uniquifier=<UUID>) failed with transport type CT, session tag INTERCLUSTER_PALOMA, record state ACTIVE, CSM error CSM_CONNABORTED, low-level error , socket error 0, and TLS error 2113929224.
- ソースとクラスタの間でキャプチャされたパケットトレースで多数のフラグメントパケットが報告される