Performing maintenance on nodes of a cluster

[duanebachi1]

Good afternoon experts,

We have a six node (3 HA pair) cluster. Our cluster is relatively  simple in terms of how we use them. All SAN via FCoE. We want to perform maintenance on one of the HA pairs because we want to relocate to a different rack. We already accept the outage for any data hosted on the HA-pair.

Is there any thing I should watch out for when I shutdown this HA pair? Anything special I should take into consideration to make sure that the cluster and the remaining HA pairs continues to function?

Thanks,

DB

[SeanHatfield · Accepted solution]

Move epsilon before you start.

Send an ASUP

Run Config Advisor 

Ensure that none of your SVM root volumes are on that HA pair

Label everything, especially the FCoE links to make sure they get reconnected correctly.

Inhibit takeover when you shut them down

After you bring them back up, verify FCoE/ALUA/MPIO

Run ConfigAdvisor again

Do a Takeover/Giveback test in both directions

Send another ASUP

[duanebachi1]

Thanks Sean,

I verified that epsilon is not on the affected nodes. If it is, how do one go about moving this to another node?

I verified that the root volumes of SVMs are not on the affected aggregates. WE do have the data protection for the SVM root volumes, is it not the purpose of that is to have redundancy?

[TMADOCTHOMAS]

Moving epsilon:

cluster modify -node <current_node> -epsilon false
cluster modify -node <new_node> -epsilon true

As for the load sharing, I'm asking about that as well in another thread (https://community.netapp.com/t5/Data-ONTAP-Discussions/Temporary-Shut-Down-of-Two-Cluster-Nodes/m-p/134519#M29459) because I need to do the same thing. Hopefully we can get an answer!