Thanks to a reader of the blog and another friend of mine who is sticking closer to the pfsense topic than me, I got a hint about the latest snapshot of pfsense which now has the Hyper-V integration services built in.
You can get an overview of the BSD Integration Services (BIS) at the official FreeBSD wiki.
Now I have a better view and I can see that one other cluster have the described problem.
With VMM you can get the message if you try to move, start up a VM:
Error (12711) VMM cannot complete the WMI operation on the server (hyper-v-03.contoso.com) because of an error: [MSCluster_Resource.Name="SCVMM SOF-Prev-02"] […] The requested operation can not be completed because a resource has locked status (0x1748) […]
Failover Cluster Manager will say “The resource $VM is locked.”
From roughly around 70 VMs almost a third can’t be moved, started up or something else.
Really painful is the fact that you can’t do anything with the Hyper-V Host itself.
Trying to get ’em in paused mode for example.
Dynamic Optimizing is getting crazy as well. 😉
Applying the Hotfix seems to be working.
I struggled with a Live Backup issue on Ubuntu 2014 LTS a while ago as ralish from Stockholm DevOps describes here.
The problem was a broken installer meta package and ends up in a missing kvp deamon if you tried to install it on Ubuntu.
This daemon is used to interact between Host and VM to get a consistent backup state during snapshot creation on the Hyper-V Host.
So at the end, Backups without kvp can be risky due to a service interruption of the VM workload(s).
A screenshot from the DPM 2012 R2 Backup console shows the offline state.
In the meantime, the bug got fixed, Live Backups without service interruption can be taken, the sun is shining again and so it’s time to wrap some things up and give you a short overview about Hyper-V and common Linux Distributions.
For the device-specific module (DSM) named Microsoft DSM, versions do not match between node CL-N01.contoso.com and node CL-N02.contoso.com. The Cluster validation of a couple of Servers 2012 R2 failed with VMM and the Failover Cluster Manager as well. Continue reading →
As Ben Armstrong wrote a couple of years ago “There is a lot of confusion about how time synchronization works in Hyper-V” …
To demystify that and get the things clear, he wrote an artcile which was updated frequently.
This stuff is very useful to know and to avoid any troubles. It’s a Kind of Basic knowledge and understanding.
Read it soon If you already missed it. 😉
In some cases it’s good to replicate a virtual machine in and outside a clustered system.
I tried this the very first time but after setting up the replication broker role in the failover cluster manager I got an error.
After researching a while whats going on I got a solution.
The Cluster Role couldn’t write a computer object.
Usually you set up an OU where the Hyper-V Servers belongs to and there is the Cluster Role (actually a computer account created during the installation) as well.
The Cluster (machine account) has to get writable access to this OU and after that… The Replica Broker is working. 🙂
I tried to convert an old Server 2003 R2 with Citrix on top today but the Virtual Machine Manager couldn’t grab and convert the HDDs.
I got different errors as I also tried it again via offline conversion. The WinPE couldn’t get any informations about the HDD (RAID 10 on LSI SCSI Controller).
Finally I cracked it as I backed up the whole machine and recovered it inside a Virtual Machine.
My experiences with old Systems an on/offline P2V conversions aren’t very well.
Within the last weeks I tried unsuccesfully to virtualize some old Windows XP and Server 2003 (R2) Systems.
The lucky thing was always to catch them via regular backup routines and recover them as I wrote above.