Hi, Posted in the Google Group Isilon-users also but no response.
I'm migrating my Celerra content to Isilon and need to support moving the existing Celerra IP's to Isilon, due to time constraints in the migration I can't wait for users to re-configure connections to use SmartConnect.
Each IP is unique in a VLAN. So I setup my pools of one IP 192.168.1.21-21 and group net specific to a VLAN, added the Agg nics from each but no matter what type of failover I set the Isilon does not move that IP address when I down a node.
Is there something I'm missing or not reading or does it just not work this way ?
Its quiet different in Isilon, if you like to take advantage of IP redundancy, you should have atleast a few IP addresses per VLAN, assign to a particular IP pool. As many VLANs you have, its good to have those many IP pools. And depending on your traffic (NFS or SMB) you can configure these IP pools as dynamic or static.
What is your Allocation Method in that Pool? Static or Dynamic? It should be set to Dynamic if you want the IP to failover to another node if the original node with the IP was down/rebooted for any reason.
Looks like I have it setup right
isiCluster-01# isi network pools view Hosting.####
Access Zone: Hosting
Allocation Method: dynamic
Aggregation Mode: failover
SC Suspended Nodes: -
Description: Pool VLAN###
Ifaces: 1:10gige-agg-1, 2:10gige-agg-1, 3:10gige-agg-1
IP Ranges: 10.189.135.124-10.189.135.124
Rebalance Policy: auto
SC Auto Unsuspend Delay: 0
SC Connect Policy: round_robin
SC DNS Zone Aliases: -
SC Failover Policy: round_robin
SC Subnet: -
SC Ttl: 0
Static Routes: -
You can append that single IP to the pool. With a dynamic allocation method, if a node goes down that IP would failover to another node, but keep in mind clients won't be happy about it, especially SMB/CIFS clients, because SMB is a stateful protocol, and session state is not maintained between nodes. If it's NFSv3 clients, that's far less of an issue because it's a stateless protocol, and NLM locks are synced between nodes.
Of course it should go without saying, but each VLAN is a different subnet, so you need to add a unique subnet, add the VLAN Tag number, so that traffic is handled correctly, and then create the pool in that VLAN. But keep in mind that your clients will always be talking to just a node when you do this. Sometimes better to cut them off at the knees when you've been telling them for years to use the DNS name.
Also when you down the interface on the VNX/Celerra, and remove the CIFS server, you'll want to delete the machine account from AD. Then set the smart connect zone name on the target to the FQDN of the source CIFS server, then create an NS delegation to the Isilon cluster's SSIP, or the SSIP for this subnet, depending on what smartconnect service subnet you use. Then you'll need to create SPNs inside the Isilon cluster's machine account in AD (assuming there is some SMB here) for the cutover CIFS server's name. Luckily you can do this from the Isilon cluster with:
'isi auth ads spn check/repair'. That syntax varies a little bit based upon your version of OneFS, so check the CLI guide for your version.
Principal Pre Sales Consultant, Datadobi