at one of my customers I faced the VIX unfreeze timeout error.
Our support sent me some backgound Information that brought me to right solution at this customer.
Customer implemented a fresh Exchange 2010 install at a new storage System with enough disks.
Added one new user and tried to backup => VIX error
We checked IO load => nothing with a Peak of 300 IO/s => Latency under 5 MS
So Disk Performance itself was not the problem.
Support sent me Information that describes the process as follow:
As for the VSS errors:
The issue itself is related to 20 second built-in hard-coded timeout in MS Exchange VSS writer.
Here is how backup process works:
2.1. initiate VSS freeze of VM, there are two ways. admin share(b) and VIX(a)
a) we connect to VC over VIX(tcp 443,902), VC passes through our agent to ESX/ESXi, ESX/ESXi passes data inside VM, and our agent is being executed via VMware tools.
b) we connect to admin share and push our agent via SMBFS/CIFS protocol after deploying of our agent, we initiate VSS freeze of the VM at the point when VM is frozen, countdown begins, Exchange VSS writer can't stay frozen longer than 20 seconds. this is by design of Microsoft.
http://msdn.microsoft.com/en-us/library ... s.85).aspx
2.2. we go to vSphere and create a VMware snapshot of the VM
2.3. when VMware snapshot is created we go back to VM and unfreeze the VM
2.4. then data transfer itself and then snapshot commit
So, between step 2.1(when countdown begins) and step 2.3 we have only 20 seconds. If Veeam doesn't freeze the VM in timely manner, Windows VSS subsystem does it by itself. Hence you get unfreeze error.
Snapshot creation for the VM should take only couple of seconds! If it takes any longer you won't make it in 20 seconds window.
Also I should say that operation via VIX are slower than operations through admin share.
Also operations through ESX/ESXi server are usually faster than operations through VC server.
So a Snapshot creation time under 20 seconds (20 seconds - Overhead to be trur) is also key to solve the vix error.
At this customer the new ESX Server was added to an existing vcenter that was under heavy load.
In this tread someone reported that he solved the Problem by backup the VM directly over a single ESXi Server Veeam Connection (not though vcenter). That reduced snapshot time and... backup worked.
At the customer we opened a VMware ticket to investigate why the snapshot creation Needs so much time.
Support also gave the tip that networkless VSS processing can reduce VSS processing time.
With 6.1 HF1 you can manipulate VSS Connection order by adding a registry key (networkless first)
http://forums.veeam.com/viewtopic.php?f ... 207#p60260
Here you need to use an predefined user "Administrator" for our VSS "Exchangeservername\Administrator" or "Domain\Administrator". An other member of Administrator Group are not valid because of MS restrictions. (you can edit these in the Group policies if needed).
Also additionally helpful can be to disable UAC
Because of (Disk)Performance and Snapshot Performance schedule the Exchange Backup Job separately when no other backup or Exchange maintenance Job runs. Check also other maintenance Windows (Virus Scanner, Windows updates,..)
VIX unfreeze error => Check Disk Performance => Check Snapshot creation time under 20 seconds => Check the other stuff described in this thread.