Forgot to mention:  TR-4015 talks about intercluster LIFS and includes failover groups.

 

From: Learmonth, Peter
Sent: Tuesday, January 13, 2015 6:10 PM
To: 'Iluhes'
Cc: Toasters
Subject: RE: cmod failover group

 

Yes, you definitely want failover groups with VLAN ports to keep LIFs within their respective VLANs.  Some examples:

eadrax::> failover-groups show

  (network interface failover-groups show)

                                  Failover

Vserver          Group            Targets

---------------- ---------------- --------------------------------------------

eadrax

                 eadrax-01_intercluster

                                  eadrax-01:e0a

                 eadrax-02_intercluster

                                  eadrax-02:e0a

                 eadrax-03_intercluster

                                  eadrax-03:e0a

                 eadrax-04_intercluster

                                  eadrax-04:e0a

                 lbt_test

                                  eadrax-04:e1b-101, eadrax-04:e2b-101,

                                  eadrax-01:e1b-101, eadrax-01:e2b-101,

                                  eadrax-02:e1b-101, eadrax-02:e2b-101,

                                  eadrax-03:e1b-101, eadrax-03:e2b-101

                 pubnet172

                                  eadrax-04:e0a, eadrax-01:e0a, eadrax-02:e0a,

                                  eadrax-03:e0a

                 stgnet42

                                  eadrax-04:a0b-42, eadrax-01:a0b-42,

                                  eadrax-02:a0b-42, eadrax-03:a0b-42

                 vlan103

                                  eadrax-04:e1b-103, eadrax-04:e2b-103,

                                  eadrax-01:e1b-103, eadrax-01:e2b-103,

                                  eadrax-02:e1b-103, eadrax-02:e2b-103,

                                  eadrax-03:e1b-103, eadrax-03:e2b-103

 

Failover groups for ICL would be similar to mgmt LIF failover groups – the failover group would consist of ports on a single node.  SnapMirror uses an ICL LIF on the node hosting the volume being replicated and will not replicate via LIF on a different node. 

 

Peter

 

From: Iluhes [mailto:iluhes@yahoo.com]
Sent: Tuesday, January 13, 2015 4:10 PM
To: Learmonth, Peter
Cc: Toasters
Subject: Re: cmod failover group

 

thanks Peter, so it looks like I created with gui tool is toast. I have to modify failover groups for all lifs that are using vlan interfaces. What about intercluster lifs? They need a second node in failover group or same story as mgmt interfaces?


On Jan 13, 2015, at 5:07 PM, "Learmonth, Peter" <Peter.Learmonth@netapp.com> wrote:

 

Failover groups are not created automatically.  It is very tricky to determine programmatically that ports are really on the same network.  Although it would be sane to use the same port on each node for the same purpose, it is not required.  Also, we wouldn’t know that two ports on the same node are on the same network.

 

There is no value in node management ports failing over to another node if the first node goes down.  If it’s down, you can’t manage it anyway, so no point in having them fail over.  However, it makes certain sense to have a failover group within the node, if you have multiple ports on the mgmt network.  If you have a problem with a port or cable, the LIF would fail over and you could still manage the node.  Your command output doesn’t show multiple ports on the mgmt net, so the failover groups don’t do anything.

 

Your cluster mgmt LIF should have a failover group with a port on each node.

 

Share and enjoy!

 

Peter

 

From: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] On Behalf Of Iluhes
Sent: Tuesday, January 13, 2015 2:44 PM
To: Toasters
Subject: cmod failover group

 

Hi Folks,

Do the failover group get generated by netapp intial install?

Somehow I have a group call "0ba_fg" on the node that is called 06b. Looks like someone made a typo.

But most importantly: It look like the policy for node mgmt IP only have 1 node in it, is that right? Should the second node be in the policy for mgmt1 for each node?

the way it is setup now, will it failover?

 

         Logical         Home                  Failover        Failover

Vserver  Interface       Node:Port             Policy          Group

-------- --------------- --------------------- --------------- ---------------

06a

         clus1           06a:e0a           nextavail       system-defined

         clus2           06a:e0c           nextavail       system-defined

         mgmt1           06a:e0M           nextavail       06a_fg

                         Failover Targets: 06a:e0M

06b

         clus1           06b:e0a           nextavail       system-defined

         clus2           06b:e0c           nextavail       system-defined

         mgmt1           06b:e0M           nextavail       0ba_fg   

 

 

Failover

Group               Node              Port

------------------- ----------------- ----------

06a_fg

                    06a           e0M

0ba_fg

                    06b           e0M