cluster peering show availablity pending

[FelixZhou]

we are on ONTAP 9.7 on both source and destination AFF330 servers. Have snapmirror in place and working fine. After we added new nodes to the cluster, we found we couldn't set up snapmirror for new volumes on same SVM. NetApp claimed we have to config the intercluster connections on new nodes to make the cluster peering working again. Can someone share your experience?

Currently the #cluster peer show availability pending.

[TMACMD]

Correct. SnapMirror is a Full-Mesh system. Each node must be able to communicate with every other node in the cluster peer.

After you added nodes (and the appropriate intercluster LIFs), you need to use the command:

cluster peer modify -peer-addrs <local_ip1>,<local_ip2>,<local_ip3>,<local_ip4>

If you added nodes to both sides, run  the appropriate command on both sides

(you may need a -cluster <clusname> on there too)

Then check it with a "cluster peer ping -originating-node <node-01> -destination-cluster <dest-clus>"

It should be able to ping every node in the cluster at both ends

[FelixZhou]

If we don't want any replication on new nodes, is there any way to skip the intercluster setup on new nodes to maintain the peering to work.

looks after added new nodes, the existing snapmirrors still work but we are not able to set up snapmirror for new volumes.

[SpindleNinja]

I don't think so.  just don't use the new nodes as locations for the DP volumes.     

But the IPs need to be in the config. 

[FelixZhou]

we don't have network set up yet, can we give IPs for intercluster on new nodes, leave them un-connected to make the cluster peering to work again?

thanks.

[SpindleNinja]

Every node on every cluster needs a ICL IP.   So if you have a 4 node peering to a 2 node.   the 4 node needs to know of the 2 on the remote end.  and the 2 node needs to know the 4 node's IPs.