• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 2774
  • Last Modified:

Hyper-v Boot failure. Reboot and Select proper Boot device

We have a 4 node Hyper-V cluster - I shut the vm's all down and powered off all the nodes last week for maintenance to the Blade Chassis.  I powered the chassis and nodes back up, and powered on all vm's without issue.  (FYI - the SAN was not shut down).  Since then, a few VM's will all of a sudden go offline (yet they are still running) and connecting to the console of the vm shows the error: "Boot failure. Reboot and Select proper Boot device or Insert Boot Media in selected Boot Device".  If i hit ctl+alt+del it will reboot and say the same thing, if I reset the vm it does the same thing, but if I POWER OFF the vm then power it back on, it will boot up just fine.

All vm's on this cluster node access the same SAN, the same LUN, so I don't think it's a SAN connectivity issue, even the Hyper-V Node itself boots from the SAN.  

Why would these VM's intermittently stop or reboot to no disk like this?  Any ideas on where to start looking would be greatly appreciated.

Thanks!
BOOT-FAILURE-ON-HYPER-V.PNG
0
Lambton
Asked:
Lambton
  • 4
  • 2
3 Solutions
 
Cliff GaliherCommented:
Sounds like you configured your shared storage incorrectly. NTFS doesn't support being shared so your host nodes need separate LUNs, and your guests either also need separate LUNs or need to sit on CSVFS, not NTFS. The behavior you describe is typical when a conflict occurs from a shared LUN when two NTFS accesses attempt to occur simultaneously.
0
 
LambtonAuthor Commented:
Sorry I wasn't clear enough.

The VM's all boot from/access a single CSVFS
The hosts all have their own boot LUNs on the same SAN
0
 
Cliff GaliherCommented:
Your coordinator node may be having issues (workload perhaps.) i'd move it to another node and see if the problem persists.
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.

 
LambtonAuthor Commented:
Thanks Cliff; I've moved the coordinator for that CSV to another node - I'll let you know how it goes.
0
 
Philip ElderTechnical Architect - HA/Compute/StorageCommented:
In Failover Cluster Manager there should be a series of event log entries that should explain what exactly is happening to the VM's VHDX file. Please post those.
0
 
LambtonAuthor Commented:
Here's the events - the issues have stopped for a moment, since the coordinator was moved, however I think I may have found (after viewing these events) that what cliff said about the ntfs thing may have has something to do with it - I found the common thing with all the VM's that were doing this is that they were all on the same host, and they all were configured to access the same .iso file.  (C:\Windows\system32\vmguest.iso) - you'd think this would be OK to have more than one machine configured with the guest services iso, but I'm unsure, maybe that's what the coordinator does?

Either way, have a look through the events and let me know you're thoughts.  (change the files inside from docx to evtx)

thanks!
Steve
Hyper-v-events-change-to-evtx.zip
0
 
LambtonAuthor Commented:
Well this issue has seemed to stop - we removed the cd-rom access to the vmguest.iso and moved the coordinator to a different node. I'm not sure what one fixed the issue, but I'm going to say it was "NTFS doesn't support being shared", and it could be the coordinator that wasn't taking care of that correctly?  Either way it's fixed - thanks everyone!

/Steve
0

Featured Post

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.

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