It's Cluster Mode : 9.2, we are planning to change all the IP addresses asscoated with all the LIF's including Management LIF's and SP IP's as well, we need a proceedure and steps to do this and would like to know whether this is a disrtuptive or non-disruptive.
For Data LIFs, planning and downtime will be required.
Non-Disruptive fail-over is designed to work when you are simply moving the 'IP/LIF' to another Physical Node/Port (Where logical Identity/address remains the same but physical resources (node/port) are changed) .
In your case, you are changing Identtity/IP , hence down-time cannot be avoided.
As you wish to chane the Data LIfs IP: It could be for CIFS/NFS or iSCSI.
Data LIF: 1) Depedning upon which LIF(protocol wise) is serving which application, go with least sensitive to being with.
2) For example, if I have a following LIF on Node-01/e0c serving CIFS for SVM1: svm1_cifs_lif with IP:192.168.0.10/24 I will create another LIF on Node-01/e0c : svm1_cifs_lif2 with IP:192.168.0.20/24 [Or whatever IP/subnet]
3) During downtime window : Terminate CIFS, use network interface modify command on SVM to point to new IP, change the DNS records in AD to point to new IP. Restart the CIFS.
4) Similarly any applications that have been exported via NFS IP, perform the same (create a new data LIF/IP). During downtime [Stop any applicaitons] , unmoun share, modify svm lif to point to new LIF/IP, re-mount the share.
5) iSCI: Create new LIF with new IPs on the iSCSI SVM, Disconnect iSCSI sessions on the Clinet, modify the svm LIF to new LIF/IP. Discover the Target with new IP.
Management LIF: For Management LIFs, there is no impact on the Data serving LIFs.
For example: To provision a new cluster-mgmt LIF IP.
Create a new lif with role cluster-mgmt with the new IP address: cluster::*> network interface create -vserver cluster -lif cluster_mgmt_new -role cluster-mgmt -home-node <node_name> -home-port
# Verify if you are able to ping the gateway from the newly created LIF/IP. # Confirm if you are able to access this new cluster management LIF via putty from you pc. # Delete the old cluster management LIF.
For any management network IP change, make sure you re-discover them in OCUM and also if there are any NDMP Jobs running in node-scoped they will need to be stopped/killed. If you are using backup software, then you will have to update/re-add the Cluster Mgmt LIF into your Array list, this is applicable to CommVault backup software.
You can do in this order: 1) Cluster mgmt first 2) Node mgmt next 3) Data Lif next [with least priority/sensitive first, then next]
Calling all Network experts @netapp forum for any corrections on the above procedure.
To add to what Ontapforum said... If you have something like VMware datastores, you can "re-IP" non-disruptivitly. But the process is not so much as a "re-IP" but more a build out a new VLAN for the NFS/iSCSI datastores, and present out new datastores on using the new vlan. VMs would be move host side via Storage vMotion (or whatever process your hypervisor uses)