Help! Hyper-V with Failover clustering has VM Domain Controller not starting... cannot access Failover cluster..??
-
Tuesday, May 15, 2012 8:44 AM
Hi.
I have made a serious mistake and need some help.. I have a Hyper-V setup using Failover Clusters. The problem is my domain controller was also virtualized on the Failover cluster. The VMs shut down and now the VH can't access the Failover Cluster, because no domain controller (yes, now I see a big reason for physical, separate domain controller!) What can I do to get my VM servers operational again?
Is there a way to access the Failover Cluster disc without the domain controller? Can I access it as a normal iscsi disk mapped to a drive letter somehow? (If so, how?) And then start my DC VM inside the host?
Is there a way I can bring the Failover Cluster online without a domain controller?
Any quick help would be very appreciated!!
All Replies
-
Tuesday, May 15, 2012 9:26 AM
Ok, i think first you need to bring the PDC up and running ASAP on another host, more details should be necessary anyway if you have a standalone host that is perfect. you should bring online the SAN disk (where the DC is located) from disk manager and copy vhd away.
Reconfigure IT identical on the Standalone host if you have it or you can even configure it as a non high-available VM on one of the cluster nodes.
Then, when the PDC is up again you could bring online cluster name and other resources and plan the installation of a physical PDC
/Mat
-
Tuesday, May 15, 2012 10:31 AM
So can I connect to the iscsi volume that is in the failover cluster, and access it from another computer like a normal disk, even though it is part of failover cluster?
-
Tuesday, May 15, 2012 10:32 AMThe other thing I have is an old backup of domain controller, but I think it is having issues because it is too old, and all the cluster info isn't up-to-date on it or something..?
-
Tuesday, May 15, 2012 11:03 AM
Stop the cluster service on one node if is running, then if you can find the iscsi volume in disk manager should it be offline. Put it online if you can and this should let you reach the vhd.
Then you could copy all files via robo and link (not import) the DC on another Standalone hyper-V.
Pay attention to every virtual DC considerations:
/Mat
-
Tuesday, May 15, 2012 11:38 AM
Well, I have learnt a big lesson here!!!
But things are now looking more positive. I had an old domain controller, and was able to use that to get the clusters visible again (although had to use NETDOM RESETPWD to put VH servers back into old domain, and had to repair the AD for cluster name). Now seeing if I can get everything back online again after copying newer domain controller etc...
-
Tuesday, May 15, 2012 3:31 PM All up and running again now.. phew! Thanks to everyone for the help. I still have fix the final setup a bit more to make it much safer for the future...
- Marked As Answer by Vincent HuModerator Thursday, May 17, 2012 3:09 PM
-
Friday, May 18, 2012 1:00 AM
BTW If it helps anyone, after copying the proper DCs to storage outside the failover cluster, I put them online again but not connected to network. Then I unplugged the old temporary DC, connected the proper DCs to network, and used NETDOM RESETPWD again (and repair AD for Failover Cluster Name), and everything came online again ok, and back to normal.
These are some of the articles I found useful...
About NETDOM RESETPWD:
http://support.microsoft.com/kb/325850
Repair AD Cluster Name:
http://support.microsoft.com/kb/950805
Recover VMs:
http://eniackb.blogspot.com.au/2009/05/how-to-recover-hyper-v-virtual-machine.html
http://social.technet.microsoft.com/wiki/contents/articles/349.hyper-v-how-to-find-vm-files.aspx
About DCs and Failover Clusters:
http://www.ms4u.info/2011/05/why-you-should-not-running-domain.html
Seizing/Transferring FSMO roles:
http://support.microsoft.com/kb/255504
http://support.microsoft.com/default.aspx?kbid=324801&product=winsvr2003
-
Friday, May 18, 2012 1:03 AMJust to let you know, I tried to connect to iscsi volume on another computer, but it could not read it all (said it was RAW). Although I didn't first make sure cluster service was stopped on the two servers... (However, because I had an old backup of domain controller, I was able to revive the FC enough to copy off the VHDs I needed...)
.png)
