Sunday, 15 October 2017

Using InterCluster LIFs for NDMP Backup and not breaking Cluster Peering* (UPDATED)

*For Cluster Peering to work 100%, a full mesh is required.

If you’ve got a requirement to configure InterCluster LIFs for NDMP backup, and these IP addresses will not be routable on your Cluster Peering network, you’ll need to move the InterCluster LIFs you use for replication into a REPLICATION IPspace (the backup Intercluster LIFs need to be in the same IPspace as your SVM, which usually is going to be default.)

Here’s an example!

Part 1: Preliminary Information

cluster2::> cluster peer show
Peer Cluster Availability
------------ ------------
cluster1     Available  

cluster2::> snapmirror show

Source          Destination Mirror  Relationship         
Path       Type Path        State   Status       Healthy
---------- ---- ----------- ------- ------------ -------
svm1:alpha  DP  svm3:alpha_dr
                            Snapmirrored
                                    Idle         true

cluster2::> net int show -role intercluster
            Logical    Network            Current       Current
Vserver     Interface  Address/Mask       Node          Port  
----------- ---------- ------------------ ------------- -------
cluster2
            intercluster2
                       192.168.0.233/24   cluster2-01   e0g   

cluster2::> ipspace show
IPspace Vserver List   Broadcast Domains
------- -------------- -----------------
Cluster
        Cluster        Cluster
Default
        cluster2, svm3 Default


Part 2: Moving the Intercluster LIF to a REPLICATION IPspace

Note: Everything is done on cluster2


cluster2::> snapmirror quiesce *
cluster2::> net int modify -lif intercluster2 -vserver cluster2 -status-admin down
cluster2::> net int delete -lif intercluster2 -vserver cluster2
cluster2::> broadcast-domain remove-ports -broadcast-domain Default -ports cluster2-01:e0g
cluster2::> ipspace create -ipspace REPLICATION
cluster2::> ipspace show

IPspace     Vserver List   Broadcast Domains
----------- -------------- -----------------
Cluster
            Cluster        Cluster
Default
            cluster2, svm3 Default
REPLICATION
            REPLICATION    REPLICATION

cluster2::> broadcast-domain create -ipspace REPLICATION -broadcast-domain REPLICATION -ports cluster2-01:e0g -mtu 1500
cluster2::> net int create -role intercluster -home-port e0g -home-node cluster2-01 -vserver REPLICATION -lif intercluster2 -address 192.168.0.233 -netmask 255.255.255.0
cluster2::> cluster peer modify -ipspace REPLICATION -cluster cluster1
cluster2::> snapmirror resume *


Note: You may need to add a route to the REPLICATION vserver to get this to work.

Part 3: Verifying All’s Working


cluster2::> snapmirror update *
cluster2::> snapmirror show

Source           Destination Mirror  Relationship       
Path        Type Path        State   Status       Healthy
----------- ---- ----------- ------- ------------ -------
svm1:alpha  DP   svm3:alpha_dr
                            Snapmirrored
                                     Idle         true

cluster2::> cluster peer show

Peer Cluster Availability
------------ ------------
cluster1     Available


Part 4: Creating InterCluster LIFs for NDMP Backup Purposes

Now you’ve moved your intercluster LIFs used by replication into their own IPspace, you can go ahead and create the intercluster LIFs for NDMP backup in your Default (or whatever) IPspace, without fear of breaking Cluster Peering and your SnapMirror replications.

Further reading:

Image: Using IPSpaces for Cluster Peering

No comments:

Post a Comment