Want to protect your cyber security and still get fast solutions? Ask a secure question today.Go Premium

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 3072
  • Last Modified:

Move VM's out of cluster to standalone Host

Hello Experts

we are planning to re-build our DELL MD 300 storage server and 2 Nodes connected to it.
below is current setup.

2 Dell Servers connected to DELL MD 3000 Storage, forming a Failover cluster and on hyper V we got 2 virtual machines running. one is domain controller and another is exchange server.

as we want to rebuild the cluster setup along with storage. so we want to move these two VM's to a new Hyper V server which is a standalone Hyper V server with enough internal Disk space and RAM to host two VM's.
one new server I installed System Center Virtual machine manager to migrate machines from clustered hosts to standalone host.
I added new server to Virtual machine manager.  and I can see that we can add Cluster to VMM as well and that will add both cluster nodes to VMM, but my question is that is it possible to move VM's from cluster environment to a standalone Hyper V server.

idea is to move VM's to standalone server and then rebuild servers with storage and setup clustering again and at the end move back VM's to newly created Cluster.

please advise if above is possible as it is or any other steps required.

Many Thanks.
Harry
0
H-Singh
Asked:
H-Singh
  • 10
  • 4
  • 2
  • +1
1 Solution
 
Rahul AgarwalTeam LeaderCommented:
0
 
Svet PaperovIT ManagerCommented:
Basically, you don’t need to add your temporary server in the SCVMM in order to move the machines.

What I would do is to shutdown the VM, merge all snapshots, if there are any, and export the virtual machines. Then I could import them into the temporary server. When the rebuilt of the cluster is done I would perform the same procedure back.
0
 
H-SinghAuthor Commented:
Thanks Guys, i think SCVMM option is not going to work for me as trying to add Cluster to SCVMM but failed to do so ended up some storage related access errors. but that was bit faster option i was thinking off.

Exporting Machines is another option we have but am just worried about Time it will take to export each machine, as I got
DC about 250GB and Exchange almost same.

any other faster option?
also there are few snapshots with DC, am not much used to working with snapshots , DELETE snapshots? what impact snapshot will have on VM. please advise .

Thanks again
0
2017 Webroot Threat Report

MSPs: Get the facts you need to protect your clients.
The 2017 Webroot Threat Report provides a uniquely insightful global view into the analysis and discoveries made by the Webroot® Threat Intelligence Platform to provide insights on key trends and risks as seen by our users.

 
Svet PaperovIT ManagerCommented:
First, you should never do snapshots of DC. You cannot use them without breaking your AD DS. Second, snapshots in production environment are not a good idea. Snapshots make your VM running slowly.

In order to merge the snapshots you just need to delete the snapshot tree.

You could export the machines with snapshots but sometimes the import crashes, at least that was the case with Hyper-V v1 in Windows Server 2008 (before R2). So, MS suggests to merge the snapshots before exporting a VM.

Export/import is the fastest way.
0
 
H-SinghAuthor Commented:
you mean just delete the snap shot subtree from the top and then once i will shut down DC it will merge automatically and after that i can export .

anyway to be on safe side as always am just backing up everything first.

is there any other points I need to be take care of?
0
 
Svet PaperovIT ManagerCommented:
Yes, when you are doing the import you will have 4 choices:
-      Move/restore using the same unique ID
-      Move/restore using the same unique ID but duplicating the files
-      Copy the VM and create new ID
-      Copy the VM and create new ID but duplicating the files
If you perform just “Move/restore using the same unique ID” it will restore the VM but will keep the VHD in the source folder. If you do “copy”, however, you could find yourself with deactivated Windows.

I would do “Move/restore using the same unique ID but duplicating the files” – the restore will copy all files in the default locations specified in the Hyper-V Manager.

The merge of the snapshots you can perform after shutting down the VM – it’s safer.

Always, the backup first.
0
 
H-SinghAuthor Commented:
thanks i will be exporting tmmrw night, after compete seperate backups of each machine and deleting snap shots for DC,

i will export directly to new server by sharing a Drive over network so it will save import time.

I will update this topic here later on.

Regards
Harry
0
 
H-SinghAuthor Commented:
I tried export import over weekend and ran into serious troubles, Domain controllers once imported to new server were giving following errors.
"the active directory database has been restored using an unsupported method"

and complete network was going down.
luckly I didnt remove original DC VM from existing host and started that up .
having searched over internet found out that, EXPORT/IMPORT of Domain controllers is not supported.  

this is no good.
0
 
NetfloCommented:
Hi Harry,

From reading the posts so far, I would advise getting SCVMM up and running first. Then move the DC VM from the cluster to the stand alone box.

I'm assuming you've connected your MD3000 SAN in a HA mode with 2 HBA per node. When moving machines in SCVMM it will ask you where it should place the VM on the destination server, which will be a path on the local drive on the destination server.

The DC VM will be paused and transferred across using BITS, for the size of the VM I agree and would advise for this work to be done out of hours.

Can you please explain in a little more detail when attempting to Add Host in SCVMM, after authenticating and adding one of the cluster nodes, it should prompt you that this machine is part of a cluster and would you like to add the cluster.

This would be the only supported method that you have at your disposal to get the VM moved cleanly. Please take a look at the following link, especially the last comment for what you're trying to achieve: http://blogs.technet.com/b/chrad/archive/2010/05/09/wait-a-minute-virtual-machine-isn-t-highly-available-though-it-is-on-a-clustered-host.aspx

A little further reading to bring you up to speed on virtualised domain controllers: http://blog.concurrency.com/infrastructure/virtualization/should-i-virtualize-my-domain-controllers/

Hope this helps,
0
 
H-SinghAuthor Commented:
Hello Netflo

thanks for advise, Yes at very first attempt I did try to do evrything using SCVMM,

SCVMM is installed on new server where I want to move all VM's out from cluster.
as I try to add Clustered hosts in SCVMM, it does goes though add process and I can see both hosts in SCVMM, but with warning marks on them and on details says host Refresh task failed.

its something like refresh host itself to bring its details in SCVMM, but that task is not going through and i didnt really wanted to mess up things on cluster as its in production and we can't afford downtime at all.

we have two DC
DC1 : This is main original DC since day one
DC2 : we added this later on as new VM as sacondary DC but now am having errors on this as i restart this DC it says in AD error logs that "the active directory database has been restored using an unsupported method" and Net logon service is paused.

third server is Exchange.

so now AD is working fine from DC1,  and looks to me DC2 is no more helping us. what you suggest shall we demote DC2 and create new VM again as DC on new server and fire that up to keep AD running while we do move/migrate main DC from cluster to new server.

but even for that I need cluster working in SCVMM.
0
 
H-SinghAuthor Commented:
sorry below is setup.

2 DELL Server connected to MD3000 DAS using one HBA each.
Cluster is working fine I can failover to another host etc. i know physical hardware is not connected using HA method its using single HBA card and single SAS cable to connect hosts to storage.
that's also in plan as we will rebuild cluster from scracth then i will get another HBA cables as HBA cards got 2 slots in each . or may be will look at buying two more HBA cards.
0
 
Svet PaperovIT ManagerCommented:
Sorry, my bad, I wasn’t aware that it could be a problem with the DC if it is imported. Which import option did you use? Did you try “Move/restore using the same unique ID” without duplicating the files?

You could also try to manually restore the VM on your stand alone server: http://eniackb.blogspot.com/2009/05/how-to-recover-hyper-v-virtual-machine.html

As last resort, you could build a new DC on the new Hyper-V server, transfer all FSMO from the old DC and demote it.  
0
 
NetfloCommented:
I disagree with Spaperovs last comment, you don't want to say 'you could also try' without stating that this is an unsupported method. The last thing Harry wants to is to screw his network up then to call MS PSS and be told, "sorry you have an unsupported network". I have seen many times on EE that advise has been given straight off the internet and the person requesting help actions the commands or steps without a second thought.
Lets try to be proactive and advise responsibly considering there is a production network and an IT person in need.

Harry, going forward if your problematic DC is a background server and does not have any direct services attached to it (e.g. print or file server), then I would demote it. I was not aware that you had a separate DC which as long as it holds the FMSO roles, you should be okay. Heads up on configuration changes on your DHCP server pumping out the secondary DC IP, which may need to be changed.

I would tread carefully on this process and possibly run DCDIAG on your main server before proceeding, just so you are aware of any hidden issues.

Best of luck!

0
 
H-SinghAuthor Commented:
Thanks Guys

i think that's the best way  just demote DC2 now as DC1 is working fine.
and than  later on setup another DC as VM and do dcpromo on that to make secondary DC.

I have to use new server as SQL server so I will be dropping this plan of moving VM's to new host and re-setup clustered hosts. as I need to setup Ektron Eintranet for office on new server.

also am tired now wasted long weekend on this. ouch.
very soon we are geting few more new servers for CMS project then again i will try to do this. for the time i will let it be running as it is. with one DC1 and Exchange as VM and will remote problemmatic DC in few Days after DCDiag.
i have to do everything at night even on weekends as well as company here is a law firm and its in production 7 days, so no more risks I can take now.

Thanks for all your valuable suggestions again
Regards
Harry
0
 
H-SinghAuthor Commented:
Hello Guys been long time on this question and guess wot we had problem with cluster and this time there was no option instead of moving out VM's so finally I have managed to move out all VM's to TEMP Server.
I used Acronis Backup and Recovery 10 Hyper V Edition. and backed up all VM's then on restore created these VM;s on new server with Acronis. all works fine. now I will be rebuilding my cluster from scratch.

Many thanks for whoever participated in this question.

Regards
Harry
0
 
H-SinghAuthor Commented:
also secondary DC's  Netlogin issue "active directory database has been restored using unsupported method"
that issue is also fixed and its working fine now.

i found solution for that on TechNet
1. Open registry editor and make a backup.

2. Navigate to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\NTDS\Parameters

3. Find the key "DSA Not Writable" (REG_DWORD) and its value is 0x4

4. Delete "DSA Not Writable" and reboot the server

0
 
H-SinghAuthor Commented:
worked foe me
0

Featured Post

When ransomware hits your clients, what do you do?

MSPs: Endpoint security isn’t enough to prevent ransomware.
As the impact and severity of crypto ransomware attacks has grown, Webroot has fought back, not just by building a next-gen endpoint solution capable of preventing ransomware attacks but also by being a thought leader.

  • 10
  • 4
  • 2
  • +1
Tackle projects and never again get stuck behind a technical roadblock.
Join Now