Cluster peer unavailable on a 2-node Stretch MetroCluster
Applies to
- ONTAP 9.6 and later
- For ONTAP 9.5RC1 through 9.5P2 (inclusive) go here
- Cluster peering
- Cipher
- 2 node Stretch MCC
Issue
- After the MGWD was manually restarted The cluster peer became unavailable and the PSK Ciphers were missing.
- Attempts to create a cluster peer fail with the below error message:
-
Error: command failed: Using peer-address XX.XXX.XX.X: An introductory RPC to the peer address "XX.XX.XX.XX" failed to connect: RPC: Remote system error [from mgwd on node "XXXXXXXXX" (VSID: -1) to xcintro at XX.XX.XX.XX]. Verify that the peer address is correct and try again.
- These errors may also appear for an existing cluster peer relationship after an upgrade to one of the releases mentioned in the "Applies to" section of this KB.
- In a two-node MetroCluster system, the switchback after ONTAP upgrade fails with the following error:
-
Cluster_A1::*> metrocluster switchback -simulate
Error: command failed: Failed to retrieve system image status information for all
nodes. Reason: Failed to contact peer cluster "Cluster_B1" at addresses:
10.XX.XX.3, 10.XX.XX.4. RPC: Remote system error [from mgwd on node
"Node_A1" (VSID: -1) to mgwd at 10.XX.XX.3, 10.XX.XX.4]. Verify that the
nodes are healthy using the "cluster show" command and that cluster peering
is available using the "cluster peer show" command, then retry the
operation. InterclusterBrokenConnectionAlert
may be reported in health monitoring.cluster peer show
output reports the availability of the Remote Cluster asUnavailable.
cluster peer health show
is empty:-
::> cluster peer health show
This table is currently empty - No response on Port 11104:
-
::*> system node systemshell -node Node_A1 nc -zv 10.XX.XX.3 11104
nc: connect to 10.XX.XX.3 port 11104 (tcp) failed: Connection refused