2015-09-09 04:02 AM
we just added two nodes into the cluster.
Any rules should I consider to provisioning new aggregates? We have vmware/datasotre, NFS for Linux, and home. I know this is a very generic questions, are there any rules, or just follow the same way as we did to existing aggrs?
2015-09-09 07:24 AM
Some of the list of things when I add nodes below... specifically for VMware over nfs, if you vol move, then re-home and move the LIF dedicated to that datastore.. If a new datastore volume, also create a new lif for that datastore.
• Match ONTAP version
• set non-UTA2 ethernet ports to no flowcontrol
• match options settings (ip.fastpath for example)
• Setup autosupport (node specific, not a cluster setting)
• Setup date/time/ntp (pre 8.3 it is a node, not a cluster setting)
• make sure "cluster ha show" is false (>2 nodes) - should be automated
• ping cluster test
• create aggregates
• create volumes
• rename root aggregates to naming conventions
• resize vol0
• create interface groups
• create VLANs
• add interfaces to broadcast-domains / failover groups that apply for lif failover
• add interface to portsets for SAN
• modify reporting-nodes for selective lun mapping as needed.
• create lifs/set IP addresses (and add to load balancer or dns round robin / ns records)
o LIFs for Vservers (Ethernet and FCP ports)
• assign failover groups to lifs (should be no change here unless new LIFS and SVMs)
• Create additional LS/DP mirrors for SVM root to the new nodes.
• Load balance existing volumes