I have a new NetApp Ontap cluster, and I'm trying to figure out how to document procedures for using it within my organization. It has a cluster name (netapp-cm, creative, I know), and I've allocated an administrative IP address. The monitoring tools we're using seems to use the IP address as the primary key for identifying the cluster, but that seems wrong to me, as IP addresses are more or less arbitrary these days in a world full of NAT and DHCP.
When NetApp OnTap monitoring tools display multiple clusters, what do they primarily use to identify each cluster: the cluster's IP address, or the assigned cluster name?
Respuesta1
I'd say use the cluster and node names and resolve them in DNS. DNS is an infrastructure service, and if it's unreliable then you have much bigger problems. (If your DNS has a dependency on your NetApp kit, then that might be an exception, but I'd suggest you don't want such a dependency to exist in the first place).
I would, however, recommend including the IP of your admin server(s) in the hosts file on your filers. And then use the name for ssh.access and httpd.admin.access (you can use a netgroup for this as well, which is handy if you're already using netgroups).
And put the IP of the RLMs for your filers into your management station hosts file.
Respuesta2
For something like a monitoring tool, using DNS is fine and arguably preferred. That said, I'm not a NetApp expert, but if its anything like most storage, there's a cluster IP. and node IP's. I would direct your monitoring at both the cluster and individual nodes so you have a complete picture of the health status.
One other note, as for mounting storage, THAT i would argue should be done via IP. There's a likely hood that DNS may not be available in certain cases , and you wouldn't want a dependency on that for storage connectivity. Example being, what if your DNS server is located on the NetApp.