ONTAP Discussions

SnapMirror Replication Failure Since Relocation

dbakerletn
4,056 Views

Hi all,

 

I have a strange issue. I have two sites, A and B, A replicates to B using DP relationships. Since shifting the kit at site B to a new datacentre I am having trouble with SM.

 

Both nodes in the switchless cluster have an uplink to the Nexus datacentre switch, using port e0c (Twinax). Node 01 seems to have a faulty cable, which will be replaced at a later date, so in the meantime I have admin shut the port on the node.

 

Cluster peer health shows partial connectivity to the source cluster, as you would expect. The trouble I have is the SM relationships will not update, the event log indicates it is due to the Intercluster LIF on node 01 is not up.

 

As I understand the replication should still flow through node 02.

 

Am I missing something here?

 

Cheers,

Dan

1 ACCEPTED SOLUTION

robinpeter
3,967 Views

from a better design point, on a DR Cluster, intercluster LIF should have 2 ports (as IFGRP)

or you can have multiple LIF on each node (using different ports)

 

With the current situation, you might be able to update the mirror by relocating the aggr to second node.

using the folling command.

 

::> aggr relocation start -node node-01 -destination node-02 -aggregate-list aggrname

Robin.

View solution in original post

4 REPLIES 4

sgrant
4,038 Views

Hi Dan, unfortunately not. The intercluster LIFs are node specific, they will not failover to other nodes in the cluster. I'm guessing you've only the one connection per node? If possible setup a second intercluster LIF on each node to avoid this problem. Intercluster LIFs can share ports with data ports, however there are some considerations, detailed in the Data Protection Guide.

 

Cheers,

Grant.

dbakerletn
4,013 Views

Hi Grant,

 

Thanks for your response. Slightly disappointing, but it is what it is.

 

Would causing a cluster failover start replication through the secondary node's ICL?

 

Regards,

Dan

robinpeter
3,968 Views

from a better design point, on a DR Cluster, intercluster LIF should have 2 ports (as IFGRP)

or you can have multiple LIF on each node (using different ports)

 

With the current situation, you might be able to update the mirror by relocating the aggr to second node.

using the folling command.

 

::> aggr relocation start -node node-01 -destination node-02 -aggregate-list aggrname

Robin.

dbakerletn
3,929 Views

Thanks Robin,

 

Sadly the DR site is SPOF due to being a DR site, we will have to add some e0a connectivity on all nodes and provision additional ICLs.

 

Kind Regards,

Dan

Public