#Removing Cluster lif

1 messages · Page 1 of 1 (latest)

floral acorn
#

Hi,
I’m trying to remove redundant cluster lif, but it throws an error saying

Error: command failed: LIF “xyz” cannot the removed because it is required to maintain quorum on node.

Initially, I had 2 cluster interconnect ports e0a and e0b on each node and lifs were created for those.
Then, I had to use e0b as a intercluster port and had attached a new lif to it.

Now, I had 2 lifs for e0a because I had to migrate the lif of e0b initially to create a new lif for e0b.

Now, I can’t find a way to remove the redundant lif on e0a.

silk cedar
#

you should be able to remove it in advanced (or diag) mode after manual confirmation. But be aware that this is not a supported process and can lead to issues (the sitelist might need to be updated, and if you have FabricPool within the cluster, i.e. running over the Cluster backend, that might also stop working and require NetApp support to fix it)

wild jewel
#

When I get a chance I will dig up the commands. There is a diag command to run to verify which cluster lifs MUST stay.

wild jewel
#

ok. do not blame me if something goes sideways or breaks

#

set diag

#

debug cluster sitelist show

#

note the clusip1 and clusip2 for EVERY node. thesse MUST stay or really risk breaking something

#

You should be able to offline/delete the Cluster-LIFS that ARE NOT using the IPs in the ouput.

silk cedar
#

yeah, as I said, the sitelist needs to be updated if one of those LIFs is deleted. Diag-commands though, so use them at your own risk 😉

wild jewel
#

The diag command is used to confirm the site list (the cluster lifs that should NOT be deleted). You are not making changes with diag in this case

floral acorn
#

So, this is the output. I am trying to remove the clus2 LIFs which are intracluster lifs on each node because they are redundant as clus1 is already up and doing the task. Now, could you please let me know how to proceed.

wild jewel
#

You can’t remove any of those. ONTAP needs a minimum of two cluster lifs per node

#

Those are cluster lifs
Do you mean inter-cluster lifs?

silk cedar
#

why would you get rid of those? your cluster will not work correctly if you don't have all four of them

silk cedar
#

the fact that two of those are down and that all of them are on e0a makes me a bit uneasy ^^

#

I sure hope this is not a production system

floral acorn
#

This is not a production system. Also, this is still in the staging phase. This is being used to isolate bad stuff from the prod and then use forensic analysis on them. Also, 2 of them are down because initially they were on e0b which is now being used for intercluster. So, in order for me to move e0b out of Cluster broadcast domain, I had to remove the lifs, so I had migrated to port e0a. Now, e0a has 4 out of which 2 are down. I want to delete the ones which are down.

raven bluff
#

As mentioned by TMAC ONTAP wants/needs at least 2x Cluster LIFs per Node. Simply admin-enable the LIFs again which are down and call it a day (if you really have no other port available).
It does not matter if there are two Cluster LIFs on the same port (ignoring lack of redundancy) you won't need these 169.254 IPs anyway.