LIf failover test

[SrinathATT]

Cluster: m218064nclusc0001
Vserver:m2_18064_9025_NFS_vs
Group:NFS-915
Failover Targets:m218064n8060c9026:a0n-915,
                 m218064n8060c9025:a0n-915,
                 m218064n8040c9027:a0n-915,
                 m218064n8040c9028:a0n-915
a0n is the ifgrp of ports (e0e, e0g) from respective each node.

Here starts the test:

1. Take the backup if interfaces/lifs and vservers associated with that lifs.
2. Set the the home node port i.e m218064n8060c9025:a0n-915 to false. (set diag,net port mod -node <nodename> -port <portname> -up-admin false) => this is for physical por
3. Verify to access the mounted NFS share and lif status.
network interface show -vserver m2_18064_9025_NFS_vs -lif m218064_9025_nfs-915 -failover
4. Once everything good, bring the port/lif back up
set diag,net port mod -node <nodename> -port <portname> -up-admin true
5,  Revert the vserver to the home port/home ned
network interface revert -vserver vserver_name -lif lif_name


Note: In point 2 I think we need to bring down one physical port in ifgrp or I think we can bring down lif for this test…if we bring down physical port, other lifs on the same port will also get failover..which I don’t want..so I prefer to do for lif for this test
network interface modify -vserver vserver_name -lif lif_name -status-admin down
network interface modify -vserver vserver_name -lif lif_name -status-admin up

[georgevj]

---

@SrinathATT wrote:

Note: In point 2 I think we need to bring down one physical port in ifgrp or I think we can bring down lif for this test…if we bring down physical port, other lifs on the same port will also get failover..which I don’t want..so I prefer to do for lif for this test
network interface modify -vserver vserver_name -lif lif_name -status-admin down
network interface modify -vserver vserver_name -lif lif_name -status-admin up

---

Bringing down only one physical port in an aggregated interface group will not affect the LIF. Switching the LIF's administrative status  up/down also may not trigger a failover. Are you trying to see data access after a LIF failover ? You could do this by migrating the LIF manually. If you really want to see the automatic LIF falover, the best way is to pull the cables from both of the ethernet ports e0e and e0g. Obvioulsly, it will affect the LIFs hosted on that port, but thats what is going to happen in real life too.

[SrinathATT]

I am going with validation of data by Migrating lif.

Okay..Do i need to migrate lif from home node port to other HA pair node port which should be part of same failover group..

 network interface migrate -vserver vs0 -lif datalif1 -dest-node node0b -dest-port e0d

[georgevj · Accepted solution]

Yes, You can do this.

In a multi nodecluster, You can migrate the NAS data LIFs to any node any port that is part of the failover group, no only to an HA partner. But make sure that if you have any VLAN in your infrastructure, those physical ports are connected to the same VLAN. In general, make sure that the packets to and from the client can reach to the port onto which you are planning to migrate the LIF.