Server OS Upgrade

Software-based VM-centric and flash-friendly VM storage + free version

Moderators: anton (staff), art (staff), Max (staff), Anatoly (staff)

Post Reply
Elcore
Posts: 12
Joined: Tue Jan 15, 2019 2:51 pm

Tue Oct 29, 2019 4:17 pm

I am upgrading my servers from Windows Server 2008 R2 to 2012 R2. Based on all of the information that I had read, the best way to do this was to:

Remove Node 1 from cluster.
Uninstall Failover feature on node 1.
Upgrade Server OS from 2008 R2 to 2012 R2 on node 1.
Install Failover feature on node 1.
Create new cluster on node 1.

I have done this and was trying to add the disks into the storage in Failover Cluster Manager but I get a message saying that "No disks suitable for cluster disks were found." How do I add my disks into this cluster?
Serhii (staff)
Staff
Posts: 12
Joined: Mon Mar 25, 2019 3:52 pm

Wed Oct 30, 2019 9:22 am

Hello, Elcore

Could you please tell me, did you read the article at the link below?
https://www.starwindsoftware.com/resour ... r-2012-r2/

Another question, could you please tell me about the current state of the old Windows 2008 R2 cluster?
Elcore
Posts: 12
Joined: Tue Jan 15, 2019 2:51 pm

Wed Oct 30, 2019 11:34 am

Yes, I read that article. I used that as the basis when I created the environment several years ago.

The current state of the 2008 R2 cluster is that it is still running on node 2 and all VMs are running on that node. I am trying to do this upgrade of the OS on both nodes while not incurring any downtime for the business.

Is there a way to make this work?
Serhii (staff)
Staff
Posts: 12
Joined: Mon Mar 25, 2019 3:52 pm

Wed Oct 30, 2019 12:42 pm

There is one way to upgrade Failover cluster from Windows 2008 R2 to Windows 2012 R2:

1. Evict one of the nodes from the cluster, then do the clean installation, add all necessary futures;
- install StarWind
- create the new HA devices for new Windows 2012 R2 Cluster Node
- connect it in iSCSI initiator. Initialize this disks in Disk Management, as writes in this article: https://www.starwindsoftware.com/resour ... 0e77312268

2. Create new failover cluster on the Server 2012R2, then, we have two one-node failover clusters, the old cluster, and new cluster;

3. Then cope roles(VM's) from the old cluster to the new cluster;

4. Ensure the new cluster works well, then, do a clean installation on the other node:

- start the replication process from node 1 to node 2 in the StarWind management console
- connect this disks in iSCSI initiator on node 2 https://www.starwindsoftware.com/resour ... 0e77312268

5. Add to the new cluster.

Cluster in-place upgrade it not support, we need to use two clusters, the old one and the new one, then cope roles and migrate storage between the two clusters.

In summary, you need downtime. Sorry.
Elcore
Posts: 12
Joined: Tue Jan 15, 2019 2:51 pm

Wed Oct 30, 2019 1:05 pm

I currently have 2 1 node clusters, the OLD 2008 R2 cluster (Node 2) that the VMs are running on and the NEW 2012 R2 cluster that I just created (Node 1).

So I need to destroy all of the storage on Node 1 and re-create it? I was hoping to avoid that but if that is the only option then I guess that is what I have to do.

I appreciate your help thus far. I was just hoping for a more streamlined approach that would not have to include downtime.
Serhii (staff)
Staff
Posts: 12
Joined: Mon Mar 25, 2019 3:52 pm

Wed Oct 30, 2019 2:26 pm

Thank you for your answer.

I discussed this case with my colleagues, and we found one more solution:

1. Shutdown all VM's on the old 2008 R2 cluster.
2. Evict all CSV from the old cluster (don't touch the Witness disk)
3. Disconnect all targets in iSCSI initiator on the old 2008 R2 cluster node

On the new Windows 2012R2 node:

- create Witness disk for the new cluster ("Quorum" for example)


1. Connect all targets in iSCSI initiator on the new node
2. bring these disks online in Disk management console
3. Add these disks in the new cluster as ClusterVolume
4. Assign new Witness disk as Quorum
5. Import these VM's to Hyper-V on new 2012 R2 node
6. Add the VM's to Cluster.
7. Run VM's.

After that destroy the old cluster, back up the StarWind.cfg file, upgrade the operating system to Windows 2012 R2, install StarWind VSAN, stop the StarWind VSAN service, replace StarWind.cfg file with the backup file, start the StarWind VSAN service, connect targets in the iSCSI initiator, add this node to cluster.

The choice is yours.
Elcore
Posts: 12
Joined: Tue Jan 15, 2019 2:51 pm

Wed Oct 30, 2019 3:37 pm

I appreciate you finding another option. It seems like there may be a little less downtime with the 2nd option but would still require all the VMs to be taken down. Since it has been about 4 years that the OLD cluster was created, I could take this opportunity to clean some things up with the storage and simplify things some.

In order to do this, I will need to take the VMs down over a weekend in either case, so the only difference is the time to copy all of the VMs and files for our file server. The OLD cluster is still up and running so as far as the employees are concerned, it's business as usual. I will just have to work this weekend, no big deal.

I assume that I need to disconnect the targets to Node 1 in the OLD Node 2 before I delete the storage within Starwind Console on Node 1?

I greatly appreciate all the help you have provided.
Serhii (staff)
Staff
Posts: 12
Joined: Mon Mar 25, 2019 3:52 pm

Wed Oct 30, 2019 3:47 pm

Elcore
I assume that I need to disconnect the targets to Node 1 in the OLD Node 2 before I delete the storage within the StarWind Console on Node 1?
You don't need to delete storage from Node1.

What the state of StarWind VSAN now?
Elcore
Posts: 12
Joined: Tue Jan 15, 2019 2:51 pm

Wed Oct 30, 2019 3:53 pm

The state of StarWind on both nodes is that it is up and running and synchronized.

I was thinking that I might take this opportunity to change the storage disks. The way I configured it years ago, is not the way that it has been best utilized. I am running out of space on 2 of the 3 Images that I created years ago and have too much space on the 3rd Image so I was going to take this opportunity to clean that up.
Serhii (staff)
Staff
Posts: 12
Joined: Mon Mar 25, 2019 3:52 pm

Wed Oct 30, 2019 4:07 pm

Ok, then it’s better to choose option one -

1. Evict one of the nodes from the cluster, then do the clean installation, add all necessary futures;
- install StarWind
- create the new HA devices for new Windows 2012 R2 Cluster Node
- connect it in iSCSI initiator. Initialize this disks in Disk Management, as writes in this article: https://www.starwindsoftware.com/resour ... 0e77312268

2. Create new failover cluster on the Server 2012R2, then, we have two one-node failover clusters, the old cluster, and new cluster;

3. Then cope roles(VM's) from the old cluster to the new cluster(don't forget about HYPER-V import);

4. Ensure the new cluster works well, then, do a clean installation on the Win2008R2 node, install the StarWindVSAN and:

- start the replication process from node 1(Windows2012R2) to new node 2(Windows2012R2) in the StarWind management console
- connect this disks in iSCSI initiator on node 2 https://www.starwindsoftware.com/resour ... 0e77312268

5. Add this new node to the new WIn2012R2 cluster.
Elcore
Posts: 12
Joined: Tue Jan 15, 2019 2:51 pm

Wed Oct 30, 2019 5:01 pm

I appreciate all the help. The support that your team offers is 2nd to none!
Serhii (staff)
Staff
Posts: 12
Joined: Mon Mar 25, 2019 3:52 pm

Thu Oct 31, 2019 8:13 am

Glad to hear it :)

Let me know if you will need additional information.
Post Reply