Wednesday, 1 January 2014

Walkthrough: Recreating a Cluster Peer Relationship

Note: To setup an Intercluster SnapMirror Relationship see here.

We have an existing Cluster Peer relationship between clusters nac1 and nac2 (by-the-by nac1 is CDOT 8.1.3 and nac2 is CDOT 8.2,) and we have existing Intercluster SnapMirrors running from the source cluster - nac1 - to the destination cluster - nac2. We want to delete the Cluster Peer relationship and recreate it.

If we try to delete a cluster peer relationship whilst SnapMirror source volumes exist in the peered cluster, the following error is received:

nac2::> cluster peer show
nac2::> cluster peer delete nac1
Error: command failed: Unable to delete peer relationship.  Reason: A Snapmirror source exists in this cluster.

Procedure

Note: All the commands here are run on the destination.

1) Record all the existing SnapMirror relationships from the source Cluster which is going to be un-peered

nac2::> snapmirror show -source-cluster nac1

2) Record the SnapMirror schedules (these will need to be re-applied later)

nac2::> snapmirror show -source-cluster nac1 -fields schedule

3) Break the SnapMirror relationships

nac2::> snapmirror break -source-path nac1://vs1/nac1Vol -destination-path nac2://svm1/nac1VolDR

4) Delete the SnapMirror relationships

nac2::> snapmirror delete -source-path nac1://vs1/nac1Vol -destination-path nac2://svm1/nac1VolDR

5) Delete the peered Cluster

nac2::> cluster peer delete nac1
nac2::> cluster peer show

6) Recreate the Cluster peer relationship

nac2::> cluster peer create -peer-addrs 192.168.168.40 -username admin
nac2::> cluster peer show

7) Resync all the SnapMirrors

nac2::> snapmirror resync -source-path nac1://vs1/nac1Vol -destination-path nac2://svm1/nac1VolDR
nac2::> snapmirror show -source-cluster nac1

8) Reapply the SnapMirror schedules

nac2::> snapmirror modify -destination-path nac2://svm1/nac1VolDR -schedule daily

THE END!

Additional Notes

A different option - downing all Intercluster LIFs, downing all ports used for Intercluster communication (if you have dedicated ports for Intercluster traffic), then re-enabling. This is demonstrated in the following example:

Note: We only have one node here and no fault tolerance of its one intercluster LIF - so just one Intercluster LIF in this cluster!

nac2::> snapmirror show
nac2::> snapmirror quiesce -destination-path nac2://svm1/nac1VolDR
nac2::> net int show -role intercluster
nac2::> net int modify -vserver nac2-01 -lif nac2_n1_ic1 -status-admin down
nac2::> net int show
nac2::> cluster peer show
Peer Cluster Name          Cluster Serial Number Availability
-------------------------- --------------------- ---------------
nac1                       -                     Unavailable
nac2::> net port show
nac2::> set advanced
nac2::*> net port modify -node nac2-01 -port e0d -up-admin false
nac2::*> net port modify -node nac2-01 -port e0d -up-admin true
nac2::*>  net int modify -vserver nac2-01 -lif nac2_n1_ic1 -status-admin up
nac2::*> cluster peer show
Peer Cluster Name          Cluster Serial Number Availability
-------------------------- --------------------- ---------------
nac1                       1-80-000008           Available
nac2::*> snapmirror resume -destination-path nac2://svm1/nac1VolDR

Image: The obligatory image

No comments:

Post a Comment