How to migrate Windows Server 2008 R2 file server cluster to Windows Server 2012 R2?
Hi,
Currently, we are maintaining two Windows Server 2008 R2 file clusters, i.e. staff and students. Now, we want to upgrade/migrate them to Windows Server 2012 R2. Our storage is SAN based (fiber). Its shares are mapped to file servers.
Now, in regards to planning, where should I start? Is there any major difference between Windows Server 2008 R2 and Windows Server 2012 R2 file clusters?
Currently, we are maintaining two Windows Server 2008 R2 file clusters, i.e. staff and students. Now, we want to upgrade/migrate them to Windows Server 2012 R2. Our storage is SAN based (fiber). Its shares are mapped to file servers.
Now, in regards to planning, where should I start? Is there any major difference between Windows Server 2008 R2 and Windows Server 2012 R2 file clusters?
Managing a cluster on 2012R2 is basically the same as before. Get used to PowerShell though, you'll get more use and info out of PowerShell than the GUI.
As for planning...
1. What SAN technology/brand are you using?
2. What does your storage utilization look like for the last 1 or 2 years?
3. What is the cost of adding storage capacity to the SAN?
4. Are there applications (exchange, SQL server, 3rd party, etc.) that use the cluster(s)?
5. Are you restricted to using your existing SAN solution?
5. How many sites do you have users in? More than 1 datacenter?
6. Can you plan down time or must the file services be continuously available?
7. Will you be restructuring the share and/or file structures?
8. How have you permissioned your shares and files structure?
8a. Direct on users or via group membership?
As for migrating...
1. How many SAN volumes are presented to the cluster?
2. Does your SAN support volume replication/duplication?
3. Do you use a backup and recovery solution?
Some of your possibilities:
- You may be able to duplicate/replicate a volume to a new LUN and present that LUN to the 2012R2 cluster.
- You use your backup/recovery system to restore the most current image of your existing clustered disk resources to the new cluster
- You could put replication/synchronizatio
Dan
As for planning...
1. What SAN technology/brand are you using?
2. What does your storage utilization look like for the last 1 or 2 years?
3. What is the cost of adding storage capacity to the SAN?
4. Are there applications (exchange, SQL server, 3rd party, etc.) that use the cluster(s)?
5. Are you restricted to using your existing SAN solution?
5. How many sites do you have users in? More than 1 datacenter?
6. Can you plan down time or must the file services be continuously available?
7. Will you be restructuring the share and/or file structures?
8. How have you permissioned your shares and files structure?
8a. Direct on users or via group membership?
As for migrating...
1. How many SAN volumes are presented to the cluster?
2. Does your SAN support volume replication/duplication?
3. Do you use a backup and recovery solution?
Some of your possibilities:
- You may be able to duplicate/replicate a volume to a new LUN and present that LUN to the 2012R2 cluster.
- You use your backup/recovery system to restore the most current image of your existing clustered disk resources to the new cluster
- You could put replication/synchronizatio
Dan
ASKER
Hi Mark,
Thanks for the reply but we are not interested in DFS because we have single forest, single site, relatively small organization therefore, DFS is not for us for sure. Currently, we are managing it via Failover Cluster Resource Manager and cluster disks....
Thanks for the reply but we are not interested in DFS because we have single forest, single site, relatively small organization therefore, DFS is not for us for sure. Currently, we are managing it via Failover Cluster Resource Manager and cluster disks....
ASKER
Hi Dan,
Thank you for your detailed reply. If I reply to the planning questions, it would be like this:
As for planning...
1. What SAN technology/brand are you using?
NetApp
2. What does your storage utilization look like for the last 1 or 2 years?
Storage utilization or adding more storage is not an issue.
3. What is the cost of adding storage capacity to the SAN?
No worries
4. Are there applications (exchange, SQL server, 3rd party, etc.) that use the cluster(s)?
Nope. Only Faculty, staff, and students and their departmental shares will be hosted on cluster.
5. Are you restricted to using your existing SAN solution?
Yes
5. How many sites do you have users in? More than 1 datacenter?
Only 1 site/datacenter. Pretty simple.
6. Can you plan down time or must the file services be continuously available?
We can plan downtime for few hours, not more than 3 hours.
7. Will you be restructuring the share and/or file structures?
For the time being, we will map the same shares.
8. How have you permissioned your shares and files structure?
Yes, NTFS permissions/ACL are pretty complex.
8a. Direct on users or via group membership?
Both
As for migrating...
1. How many SAN volumes are presented to the cluster?
As I have mentioned, there are two file clusters. Here is the detai:
X file cluster = 3 Services/applications, 2 nodes, cluster disks: 8 + 1 quoram
Y Cluster= 1 Applications/Services, 2 Nodes, Cluster Disks: 10 + 1 quoram
2. Does your SAN support volume replication/duplication?
I need to check
3. Do you use a backup and recovery solution?
Yes, IBM Tivoli/TSM
We don't use DFS
Thank you for your detailed reply. If I reply to the planning questions, it would be like this:
As for planning...
1. What SAN technology/brand are you using?
NetApp
2. What does your storage utilization look like for the last 1 or 2 years?
Storage utilization or adding more storage is not an issue.
3. What is the cost of adding storage capacity to the SAN?
No worries
4. Are there applications (exchange, SQL server, 3rd party, etc.) that use the cluster(s)?
Nope. Only Faculty, staff, and students and their departmental shares will be hosted on cluster.
5. Are you restricted to using your existing SAN solution?
Yes
5. How many sites do you have users in? More than 1 datacenter?
Only 1 site/datacenter. Pretty simple.
6. Can you plan down time or must the file services be continuously available?
We can plan downtime for few hours, not more than 3 hours.
7. Will you be restructuring the share and/or file structures?
For the time being, we will map the same shares.
8. How have you permissioned your shares and files structure?
Yes, NTFS permissions/ACL are pretty complex.
8a. Direct on users or via group membership?
Both
As for migrating...
1. How many SAN volumes are presented to the cluster?
As I have mentioned, there are two file clusters. Here is the detai:
X file cluster = 3 Services/applications, 2 nodes, cluster disks: 8 + 1 quoram
Y Cluster= 1 Applications/Services, 2 Nodes, Cluster Disks: 10 + 1 quoram
2. Does your SAN support volume replication/duplication?
I need to check
3. Do you use a backup and recovery solution?
Yes, IBM Tivoli/TSM
We don't use DFS
OK, then... your technology base is set. The existing clusters appear to be well documented.
The initial steps are pretty straight forward for each cluster:
1. acquire hardware
2. build out OSes on the cluster nodes
3. fully patch server OS
4. drop in, network connectivity for main network and cluster heartbeat
5. drop in SAN connectivity and a test LUN
6. present volume to cluster nodes.
7. run Failover Cluster validation test
8. drop test LUNs and volumes.
Then you create a new cluster, copy of the cluster roles, migrate the SAN resources over to the new cluster nodes, destroy the old cluster.
If you want to use new storage resources, then you will need to plan to either duplicate the LUNs and volumes or restore the file structures to the new cluster's storage resources.
Here is a good overview and planning resource from Microsoft:
link: https://technet.microsoft.com/en-us/library/dn530779.aspx?f=255&MSPPError=-2147217396
Dan
The initial steps are pretty straight forward for each cluster:
1. acquire hardware
2. build out OSes on the cluster nodes
3. fully patch server OS
4. drop in, network connectivity for main network and cluster heartbeat
5. drop in SAN connectivity and a test LUN
6. present volume to cluster nodes.
7. run Failover Cluster validation test
8. drop test LUNs and volumes.
Then you create a new cluster, copy of the cluster roles, migrate the SAN resources over to the new cluster nodes, destroy the old cluster.
If you want to use new storage resources, then you will need to plan to either duplicate the LUNs and volumes or restore the file structures to the new cluster's storage resources.
Here is a good overview and planning resource from Microsoft:
link: https://technet.microsoft.com/en-us/library/dn530779.aspx?f=255&MSPPError=-2147217396
Dan
Hi Dan,
Thank you for your comment. So, we are planning to add our Windows Server 2012 R2 nodes to existing Windows Server 2008 R2 failover cluster. Now, as I told you our existing failover cluster's nodes got mapped SAN LUNs. Remember, new nodes are VMWare VMs.
I have already added the roles to new Server 2012 VMs. Now what else should I do before adding them to existing cluster?
Thank you for your comment. So, we are planning to add our Windows Server 2012 R2 nodes to existing Windows Server 2008 R2 failover cluster. Now, as I told you our existing failover cluster's nodes got mapped SAN LUNs. Remember, new nodes are VMWare VMs.
I have already added the roles to new Server 2012 VMs. Now what else should I do before adding them to existing cluster?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
One way to go is using DFS.
https://technet.microsoft.com/en-us/library/cc732863%28v=ws.10%29.aspx for 2008 R2 and
https://mizitechinfo.wordpress.com/2013/08/21/step-by-step-deploy-dfs-in-windows-server-2012-r2/ AND http://blogs.technet.com/b/dfs_in_windows_2012/archive/2013/02/01/managing-dfs-in-windows-server-2012-using-hub-spoke-topology-in-a-domain-namespace.aspx for 2012
Hope that helps
Thanks
Mark