Netbackup hardware refresh on MSCS cluster

[skravi10]

Hi,

We are planned for hardware refresh for netbackup master server using MSCS cluster.

We would need less down time to move the existing catalog to new MSCS cluster, how do perform the activity. Please share the steps to perform hardware refresh for new master server.

[Michael_G_Ander · Accepted solution]

if it is just the hardware

1. Move MSCS cluster group and Netbackup cluster group to same node

2. Remove the passive node from the netbackup cluster  

3. Remove passive node from the MSCS

4. Install OS with same name on new hardware

5. Add the node to the MSCS

6. Install netbackup and choose add node to cluster

7. Move MSCS cluster group and Netbackup cluster group to the other

8. Repeat 2-6 for the second node

You might need to disable UAC and set DEP to Windows Programs only to use this approach, but it should only give a down time about 2x time for the failover

[skravi10]

Thanks for sharing the information Michael,

Our query is we dont want to disturb the existing setup incase it will take long time to complete the activity we would need to use the existing setup to avoid major impact on the existing backup infrastructure.

Instead of move the netbackup cluster and MSCS cluster to the other new node, Is it posiible to take a backup of export the netbackup cluster and MSCS cluster?

Above mentioned scenario will impact the existing setup, incase fail-over again I need to move it to existing node. Main intension is without disturbing existing setup to bringup the new hardware.

1. OS Will install with different name because we are using virtual IP reference to master server hostname.

2. Make it to ready for MSCS.

2. Catalog images and netbackup databases are configured in SAN, before starting the netabckup installation we will present the LUN to new hardward and bringup the same drive.

3.If export is possible, we need to import the netbackup cluster and MSCS cluster.

4. Incase not possible to export and import for netbackup cluster movement, please share the alternate options.

Please help us. Thanks

[sdo]

Michael's advice is the best option - or take Michael's outline plan, and add your new cluster nodes as 3rd and 4th nodes - and run on those for a few days before ejecting original nodes 1 and 2.

With your proposal you could end up with the situation where the SAN storage has been tainted somehow by the new cluster nodes and so fails to work when presented back to the original cluster - i.e. you will have had the situation where the NetBackup database has been running on a different cluster by different binaries.  It all seems a bit over-complicated.  Michael's method is the way most of us would do this type of thing.