I am migrating from a 4-node cluster to a 2-node cluster. I have a vserver dedicated to iSCSI. Right now it is being used to boot Solaris 10 zones (I know, not good).
We are doing the migration by joining the 2 nodes to the 4 node cluster, doing vol moves to the new aggrs, and then migrating the LIFS.
We have to delete the iSCSI LIFS since iSCSI LIFS cannot be migrated.
Our Unix admins are not sure how to best approach this from the Solaris perspective to eliminate or minimize downtime. We attempted to delete an iSCSI LIF and it immediately impacted the Solaris server.
I am hoping someone has some ideas or suggestions to minimize the impact other than confirming multipathing is working on the Solaris server. One thought I had would be to add iSCSI LIFs to the new nodes with new IP's and have the Unix admins add them as new targets.
Environment:
4-node 6250's moving to 2-node A700's
(8 iscsi LIFS, two per-node) (4 iscsi LIFS, two per-node)
I can provide more detail if this is not clear.
Thanks in advance!