Aggregate Relocate ownership

[maiden geode]

Hi All

We have a customer with A250 and when it was setup in a 2 node HA pair. It was always planned to buy and additional shelf and assigned that workload to the second node but it has not eventuated.

The cluster was setup with ADP there was 2 data aggregates created and they were both assigned just to a single node.

This single node is now under pretty heave load (65%) with Fabricpool being used very heavily

Now I'm looking to migrate one of the aggregates (60TB) over to the unused node.

How long would this relocate take ?

Im assuming this can be done live (out of hours) and i don't need to do anything with the volumes on the aggr/mount points etc.

https://docs.netapp.com/us-en/ontap/disks-aggregates/relocate-aggregate-ownership-task.html

any real world experience would be good

[dense jasper]

its a non-disruptive process, so no interuption to serving clients. It's pretty quick, not too different to a standard storage failover.

[maiden geode]

i guess the only other thing to factor in is we only have 1 CIFS ip which lives on Node 01 so that would mean to access the volumes on Node 02 it will route over the 25gb cluster interconnect.. which is no big deal.. Given the Data ports are only 10gb in this setup

[dense jasper]

Yeah thats fine. If that CIFS LIF is only used to access the volumes on that aggregate you can migrate the LIF as well.

Or you can create another CIFS LIF on the other node, and add a second DNS A record for the same CIFS Server. Then clients will access which LIF IP that DNS returns to them. Some will be direct to the controller with their share, some will use the cluster interconnect. This is normal for CIFS workloads

[obsidian crypt]

the A record will have to point to the same host name or you'll have to add an SPN to Kerberos