I am wondering what might be causing this problem: I can move my management LIFs around to the ports that are available in the failover groups I created, except for one node management LIF on one node, which gives me an error:
The cluster management LIF is used to manage the cluster. It can only reside on and fail over to data ports but can fail over to any data port on any of the nodes in the cluster.
In my configuration, I am using e0M to house the cluster management LIF and node management LIF, but I am also using the ifgrp ports tagged with VLAN 110 (management network) for failover purposes. I've changed the role of those ports (a1a-110 on node 1 and a2a-110 on node 2) to "node-mgmt". The setup works (except for the glitch desribed above) but I am wondering since TR-4182 mentions that cluster management LIFs should reside on and fail over to data ports that configuring the port roles to "node-mgmt" is a mistake ? Basically I want to use the VLAN 110 ports for both cluster and node management.
I believe the recommended setup is to have a dedicated port for cluster management such as e0a and e0M for node management, these should ports should NOT be on a VLAN.We have these on the same IP segment but NetApp recommends they be on the separate IP segments.
Re: failed to migrate: failed to move cluster/node-mgmt LIF.
I have now come across this issue on two cDOT setups, with one common factor: in both cases, I had changed the cluster identity (so basically changed the cluster name) without rebooting the nodes. In the second case, I came across this problem (which can be summed up as: "network interface migrate refuses to migrate a lif") and after a node reboot, the problem disappeared.
So I am now more or less convinced that this issue was caused by a cluster identity rename without performing a node reboot.