Cannot start VM after engineer ran updates on 2008 RTM server
-
2008年4月8日 下午 10:24
We were running several guest VM's on a Server 2008 Ent (RTM) server. One of our engineers last night ran updates on the server and now we cannot restart any of the VM's, one we have to get back up today to finish some testing.
The only VM we are concerned with is called SydPacSV007 and has snapshots. When we go to start this VM it errors and says...
-
"Failed to read or update Vm Configuration"
-
"Failed to initialize"
-
APPCRASH, vmwp.exe, stackhash_90b1
Any ideas, besides killing the engineer?
-
所有回覆
-
2008年4月8日 下午 11:12版主
Was the RC0 update for Hyper-V applied to the Host?
(it is not a Windows Update delivery)
-
2008年4月9日 上午 03:17
Apparently it was the RC0 updat., According to the engineer it was listed in Windows Update but not checked by default.
Yes, he checked it.
-
2008年4月9日 上午 03:33版主
If that is the case your snapshots are not easily pulled forward.
If you want to continue running RC0 then you must commit the snapshots so you can then recreate each VM as a new VM attaching to an existing VHD and then upgrading the OS within the VM (WS08 VMs need the patch applied, other VMs need the Integration Components upgraded).
Oh, and you might need to re-create your Virtual Switches.
(this is all part of migrating from the Beta to the RC)
See this blog post: http://blogs.technet.com/jhoward/archive/2008/03/19/Hyper_2D00_V-RC0-release-is-available-for-download.aspx
Be sure to read the entire post by John Howard.
Manually committing your snapshots has been mentioned here:
http://forums.microsoft.com/TechNet/ShowPost.aspx?PostID=2981469&SiteID=17
If the VMs were automatically saved (paused) when the Host was rebooted, the saved state needs to be discarded. This can be accomplished by creating a new VM or manually deleting the state files (the VSV and BIN).
Now, another option is to try to remove the patch (yes, this makes anyone cringe)
If you have the hardware build a fresh server with the same Beta release of Hyper-V that you ran before and move your VMs over to it by putting the config folders and the VHDs in the exact same places.
-
2008年4月9日 上午 05:49
Thanks for the help. I've been able to get the VM restarted by deleting the snapshots and removing the VM configuration. I then created a NEW VM and pointed it to the VHD file and it started up. I had to load the patch (KB949219 x64) in the VM first then I had to uninstall then re-install the Integrated Features so the NIC would appear in the VM.
Again, thank you for your help.
-
2008年4月9日 下午 02:59版主
Glad that your moment of panic is over.
-
2008年4月9日 下午 03:13I lost both running VM's on one of my servers after the updates also! Host and guest running licensed release versions (Select image download). "failed to read or update VM configuration: Unspecified error (0x80004005)"
Predominant error in VMMS-ADMIN log that began appearing after updates were completed: Failed to register service principal name.
Note: updates were done manually, first on each VM, with sucessful reboots prior to updating host. Guests were running at time of host reboot.
Anyone know why this happened??
Recovery...best method?
TIA
Ed -
2008年4月9日 下午 03:24版主
Since your guests were running they were suspended (by default) when the host rebooted.
It is advised tht the guests be turned off (shutdown) when the Hyper-V is updated.
Since this is the underlying hypervisor, not the WS08 parent you see in the console - incompatibilities can happen as modules and behavior are changed.
Have you tried deleting and re-registering the VMs?
If that fails, then you might want to pursue recovery.
-
2008年4月9日 下午 03:58Turns out that even a VM that was not updated and not running at host update (hasn't been run for a few days), throws the same error now.
Also, just to clarify...The Hyper-V is RC0 -
2008年4月9日 下午 04:27擁有者
-
2008年4月9日 下午 06:11Thank you sir Ben...

