Renaming a NetApp ONTAP cluster is something that I’ve been asked about more than a few times; usually it’s related to a cluster changing locations, or when you’ve done a controller migration by adding a new pair of controllers into the cluster and removed the old pair. It’s pretty straightforward and non-disruptive. It can be manually done, or even automated (but that’s another blog!)
Let’s start by looking at the cluster name (identity).
Mother::> cluster identity show Cluster UUID: 99a0be28-9999-99ea-9a99-00a098c34804 Cluster Name: Mother Cluster Serial Number: 1-80-000011 Cluster Location: Erehwon Cluster Contact: @SpindleNinja
Note that the cluster has nodes that match the naming schema.
Mother::> cluster show Node Health Eligibility --------------------- ------- ------------ Mother-01 true true Mother-02 true true 2 entries were displayed.
Let’s start to rename!
Mother::> cluster identity modify -name WOPR
Now we can see that the identity has changed.
WOPR::> cluster identity show Cluster UUID: 99a0be28-9999-99ea-9a99-00a098c34804 Cluster Name: WOPR Cluster Serial Number: 1-80-000011 Cluster Location: Erehwon Cluster Contact: @SpindleNinja
If we look at the nodes, however, they are still called by the old cluster’s name.
WOPR::> cluster show Node Health Eligibility --------------------- ------- ------------ Mother-01 true true Mother-02 true true 2 entries were displayed.
To rename them, run the “system node rename” command.
WOPR::> system node rename -node Mother-01 -newname WOPR-01 [Job 962] Job succeeded: Rename of the node "Mother-01" to "WOPR-01" is successful. WOPR::> system node rename -node Mother-02 -newname WOPR-02 [Job 963] Job succeeded: Rename of the node "Mother-02" to "WOPR-02" is successful.
Now, we can see that the names have changed.
WOPR::> cluster show Node Health Eligibility --------------------- ------- ------------ WOPR-01 true true WOPR-02 true true 2 entries were displayed.
Note: If you just want to rename to renumber the nodes, this is all you need to do: Run the “system node rename” command and just change the -XX value. The names are all, technically, arbitrary anyway.
WOPR::> system node rename -node WOPR-02 -newname WOPR-99 [Job 976] Job succeeded: Rename of the node "WOPR-02" to "WOPR-99" is successful. WOPR::> cluster show Node Health Eligibility --------------------- ------- ------------ WOPR-01 true true WOPR-99 true true 2 entries were displayed.
And we can rename it back:
WOPR::> system node rename -node WOPR-99 -newname WOPR-02 [Job 977] Job succeeded: Rename of the node "WOPR-99" to "WOPR-02" is successful. WOPR::> cluster show Node Health Eligibility --------------------- ------- ------------ WOPR-01 true true WOPR-02 true true
Let’s take a look at the rest of the parts of the cluster.
Note that the admin and node vservers (SVMs) were renamed automatically, so there is no need to do anything here.
WOPR::> vserver show Admin Operational Root Vserver Type Subtype State State Volume Aggregate ----------- ------- ---------- ---------- ----------- ---------- ---------- iSCSI data default running running iSCSI_root N1_aggr1 WOPR admin - - - - - WOPR-01 node - - - - - WOPR-02 node - - - - - 4 entries were displayed.
The network LIFs are another story. They will need to be manually renamed.
WOPR::> net int show -lif Mother* (network interface show) Logical Status Network Current Current Is Vserver Interface Admin/Oper Address/Mask Node Port Home ----------- ---------- ---------- ------------------ ------------- ------- ---- Cluster Mother-01_clus1 up/up 169.254.206.139/16 WOPR-01 e0e true Mother-01_clus2 up/up 169.254.46.169/16 WOPR-01 e0f true Mother-02_clus1 up/up 169.254.172.145/16 WOPR-02 e0e true Mother-02_clus2 up/up 169.254.214.1/16 WOPR-02 e0f true WOPR Mother-01_mgmt1 up/up 192.168.1.221/24 WOPR-01 e0M true Mother-02_mgmt1 up/up 192.168.1.222/24 WOPR-02 e0M true 6 entries were displayed.
Renaming is straight forward, just time consuming.
WOPR::> net int rename -vserver Cluster -lif Mother-01_clus1 -newname WOPR-01_clus1 (network interface rename) WOPR::> net int rename -vserver Cluster -lif Mother-01_clus2 -newname WOPR-01_clus2 (network interface rename) WOPR::> net int rename -vserver Cluster -lif Mother-02_clus1 -newname WOPR-02_clus1 (network interface rename) WOPR::> net int rename -vserver Cluster -lif Mother-02_clus2 -newname WOPR-02_clus2 (network interface rename) WOPR::> net int rename -vserver WOPR -lif Mother-01_mgmt1 -newname WOPR-01_mgmt1 (network interface rename) WOPR::> net int rename -vserver WOPR -lif Mother-02_mgmt1 -newname WOPR-02_mgmt1 (network interface rename) WOPR::> net int show -lif Mother* (network interface show) There are no entries matching your query.
(Now that looks better. )
WOPR::> net int show -lif WOPR* (network interface show) Logical Status Network Current Current Is Vserver Interface Admin/Oper Address/Mask Node Port Home ----------- ---------- ---------- ------------------ ------------- ------- ---- Cluster WOPR-01_clus1 up/up 169.254.206.139/16 WOPR-01 e0e true WOPR-01_clus2 up/up 169.254.46.169/16 WOPR-01 e0f true WOPR-02_clus1 up/up 169.254.172.145/16 WOPR-02 e0e true WOPR-02_clus2 up/up 169.254.214.1/16 WOPR-02 e0f true WOPR WOPR-01_mgmt1 up/up 192.168.1.221/24 WOPR-01 e0M true WOPR-02_mgmt1 up/up 192.168.1.222/24 WOPR-02 e0M true 6 entries were displayed.
You also might want to make sure that you don’t need to rename aggrs. I don’t rename them here in my lab, but renaming an aggr is also non-disruptive by using the “storage aggregate rename” command.
WOPR::> aggr show Aggregate Size Available Used% State #Vols Nodes RAID Status --------- -------- --------- ----- ------- ------ ---------------- ------------ N1_aggr1 3.05TB 1.98TB 35% online 23 WOPR-01 raid_dp,normal N2_aggr1 3.05TB 2.19TB 28% online 10 WOPR-02 raid_dp,normal root_aggr0_N1 368.4GB 17.85GB 95% online 1 WOPR-01 raid_dp,normal root_aggr0_N2 368.4GB 17.85GB 95% online 1 WOPR-02 raid_dp,normal 4 entries were displayed. WOPR::>
If we take a look at Active IQ Unified Manager, the name has been updated in there:
So there we have it. It’s pretty straightforward for a procedure that one might think renaming would really confuse things in a cluster.
SpindleNinja's blog 