upgrading my Hyper V SAN Cluster to Windows 2012 R2

[MrWu]

Hi all

 

We are planning to upgrade our existing Hyper V Cluster Windows 2008 R2 (3 x HP DL380G7 and 1 x P2000 G3 SAS) to Windows 2012 R2 Hyper V Cluster. What is the best way to do this? For example build a new Cluster with a 4th Spare VM Host then migrate over, then upgrade the rest of the hosts? Would there be any issues with P2000 SAN?

 

thanks as always

[gaz350b]

we have 6 hosts on our p2000 San no issues with the p2000. Only thing I can think of being an issue is time frames/uptime requirements.

 

We created a 2nd cluster shared volume as we needed to keep some vm's running and moved 1 at a time during maintenance windows for that service.

 

Migrate vm off each host, evict node from cluster, upgrade, add to new cluster.

 

Rinse and repeat

Edited July 6, 2015 by gaz350b

[CAWJames]

We did this recently, and it isnt for the feint of heart!! We have 2 R710 servers in a 2 node cluster, both with dual procs and 170 odd GB. We migrated all our VM's to one node, then broke the cluster and wiped the empty node completely(the squeaky bottom step) and built it up with 2012R2 and created a one node cluster on it. We then used the built in wizards to migrate the VM's over from one cluster to another. Bought the VM's up in the new 2012R2 cluster, wiped and rebuilt the second node, and then added it to the new cluster!!

 

To be honest the migration wizards in 2012R2 made things a lot easier.

 

We looked at the upgrade route, rather than wipe, but we felt we would be better starting from scratch, with the only downtime being the time to move the VM's over, and as we have a SAN, this didnt take long as there was actual data to move.

 

The only caveat I will add, is that make make sure your new hosts are patched completely before you start migrating, as a reboot or pending upgrades can cause gremlins. We found our VM's in limbo as I rebooted the new 2012 host in the middle of the migration progress, took a little manual jiggery to move things completely, but all good in the end.

 

James

[rich_tech]

We have a 2 node/SAN setup which was initially 2012(HP DL380e x2 and a HP P2000 iSCSI), I recently took the decision to redo the disks and changed the setup with the SAN storage when I was upgrading both of our Server OS to 2012 R2, so in essence it was a fresh installation for us. There is no straight over the top upgrade pre-2012R2 I am told, though now its a different ballgame with R2

 

I basically exported all the machines off the cluster onto our lesser quality server and kept the basic essentials running here (printing, exams, successmaker (our domain is LEA provision as is our main file server) (we could live without spiceworks and WSUS for a day or two)

 

I then destroyed the cluster and nuked both nodes, reconfigured the SAN storage for the cluster in its management and rebuilt it all from the ground up including updates, took a little over 2 working days to complete, including the importing back of VMs onto the system, only taking that long because the servers are fairly slow to reboot here and I made sure both were updated to as high as possible (in excess of about 5 minutes a time rebooting).

 

I have a list somewhere of what I actually did in the process if you want, it may help with giving a rough guide of the procedure I followed.

Edited July 7, 2015 by rich_tech

[rich_tech]

1. Remove Failover role from existing VMs and carry out an export of the VM from the host
   
2. Transfer exports to non-clustered server and start up VMs on new host
   
3. Check working Print/SuccessMaker/Exam boxes decamped to new host
   
4. Site Test of VMs with end users
   
5. Secure Multiple copies of exported vm backups to 2 x NAS devices
   
6. Run 3 Main Services physical boxes from new host – Settings altered on them to 2GB Ram and 2 Cores, going to run essential services only
   
7. Check Printing/SuccessMaker/Exam software working fine for Staff – Staff members tested printing
   
8. Revoke Veeam licences from the two Hyper-V nodes, Backup Veeam Configurations to NAS Boxes
   
9. Disconnect within windows and remove on the SAN all iSCSI host connections
   
10. Carry out a Destroy Cluster and delete the objects from the AD
    
11. Setup Quorum Disk on the raid
    
12. Reconfigure SAN vDisks within management interface using Raid6 and larger LUN setup
    
13. Reimage host machines to Datacentre 2012 R2
    
14. Rejoin Domain with both hosts
    
15. Install Failover Clustering/Desktop Experience/Hyper-V
    
16. Install and run HP Smart Update Manager
    
17. Patch machines up with HP updates – Proliant Service Pack 1
    
18. Reboot the Machines
    
19. Patch up machines to identical levels with windows updates.
    
20. Rejoin iSCSI connections to host units
    
21. Bring Online CSV/Quorum disks and format all to NTFS
    
22. Reprogram IP ranges onto the network cards on the units
    
23. Recreate Failover Cluster and Revalidate configuration
    
24. Power down cluster and hosts
    
25. Resegment Power Connections in the UPS in Server Room
    
26. Reboot Cluster Nodes
    
27. Test Failover on Cluster
    
28. Export VMs off temporary host and copy onto the cluster
    
29. Reimport VMs back onto the cluster
    
30. Reinstall Veeam onto the hosts and re-licence
    
31. Re-setup Veeam backups to NAS devices
    

 

Thats the list that I worked too when I recently redid our cluster from scratch.