• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 227
  • Last Modified:

Hyper-V VMs failing

We have 2 Hyper-V server that have been having some issues with failover.  From what I can gather the failure is because of my first Hyper-V host.  

Everything seems to depends on him.  When he goes down so does everything.  When I check the second Hyper-V host he cannot see any of the volumes to boot the VMs.  But, once the first host comes up he then can see them.

Now we have our two hosts setup in a Quorum Config of Node and Disk Majority, but for whatever reason if we lose the first host, we still lose all our VMs.

Any ideas?
0
Computech
Asked:
Computech
  • 7
  • 4
1 Solution
 
Ned RamsayNetwork Operations ManagerCommented:
This may sound really stupid (I dont know your level of knowledge) but are the the volumes actually attached to the first host so when it goes down it stops providing ISCSI to the second? Ive seen people misconfigure storage in this way.
0
 
ComputechAuthor Commented:
I have my CSVs attached through ISCSI to both hosts.  But in checking things out today, I only see the CSVs on the first host's disk manager and not on the second host.  Yet I can access the volumes in Explorer on both hosts.

I'll admit, I'm not a Hyper-V guy but a VMware guy so I'm not going to play like the expert when it comes to this.  So feel free to say what you are feeling!!
0
 
Cliff GaliherCommented:
Sounds like your witness disk doesn't have a direct path to your second node. Have you made any recent changes and did you revalidate the configuration?
0
Creating Active Directory Users from a Text File

If your organization has a need to mass-create AD user accounts, watch this video to see how its done without the need for scripting or other unnecessary complexities.

 
ComputechAuthor Commented:
We had some backup software that screwed with things a while back and we removed it a while ago as we ran into this same issue.  Do I need to revalidate the configuration everytime I make a change?  Is that the same as Validating the Cluster?  I've ran it, but excluded the volumes as it said that it would take down the volumes and I don't want to do that!
0
 
Cliff GaliherCommented:
If you make changes to the fluster itself then yes, you really should revalidate.
0
 
ComputechAuthor Commented:
Will it really take down my volumes and kill my VMs if I do the validation?
0
 
Cliff GaliherCommented:
The logic here is that you can either interrupt service in an orderly way during a maintenance window, or you can have service interrupted when node 1 fails and node 2 doesn't take over... which apparently has happened to you once already to discover the problem.

So it isn't a matter of 100% uptime, but when and how you want that downtime to occur. As always, have backups! If your cluster is not stable,  the process indeed could make things worse.
0
 
ComputechAuthor Commented:
I'm just trying to evaluate if I can run the validation on the disks right now or if I should scheduled a time when I can.

Having said that, do you think the validation will correct the error or tell me what the problem is?
0
 
Cliff GaliherCommented:
That's what the validation tool is designed to do. Of course, of things are too far out of whack, it may not be able to. In such cases, rebuilding the cluster is easier.
0
 
ComputechAuthor Commented:
Sweet, well then I'll see what I can do tonight and go from there.
0
 
ComputechAuthor Commented:
Okay, so we figured out what was going on.  Apparently both hosts should see all the volumes.  So, we found out it was an issue with the ISCSI initiator on the server.  Once we disconnected the ISCSI and reconnected it then the volumes were seen and everything started working again.
0
 
ComputechAuthor Commented:
We were able to find a solution to the issues without following the instructions provided on this question.
0

Featured Post

Has Powershell sent you back into the Stone Age?

If managing Active Directory using Windows Powershell® is making you feel like you stepped back in time, you are not alone.  For nearly 20 years, AD admins around the world have used one tool for day-to-day AD management: Hyena. Discover why.

  • 7
  • 4
Tackle projects and never again get stuck behind a technical roadblock.
Join Now