Building VMs List

[sjohnson-scu]

Why is it so SLOW!!!! With 77 VMs it usually takes nearly an hour just to build the list of VMs from the HyperV cluster... This is just crazy!! How can it be speed up?!?! With VMWare it takes less then 5 minutes to do the same for 30VMs and I BET if I had 100VMs or more the time would still be less then 10 minutes...What is going on when its building the list of VMs?

Case #00997966

[foggy]

Scott, are there probably a lot of VMs with snapshots (checkpoints) among those 77? Internally, we've seen a considerable slowdown during backup preparation stage in this case.

[sjohnson-scu]

I just checked them all and there are only 2 VMs with checkpoints...Ill be merging them for tonight run to see if it helps.

[sjohnson-scu]

I have removed the 2 checkpoints on the VMs and its still taking forever...about an hour to build the VM list.

Could this be caused by differential disks?

[Vitaliy S.]

I don't think so, as we have many VMs running differential disks in the labs. What was the response from our support team regarding this issue?

[adrianIDC]

Also me the same thing has happened Working with cluster it takes a little longer to load the list of virtual machines.

But as you take me to list 77 MV, I have a 51 and as I say , about 3/4 seconds.

But if you add host by host it takes less (a difference of a few seconds )

[sjohnson-scu]

Case #00997966

This is a failover cluster via VMM so can't add host by host....
The 2 VMs with snapshots have been removed. All differential disks have been removed... All VMs are running dynamic VHDXs for the disks...Latest MS patches installed on the host & guest.

I setup a test job with 1 single VM and it took 1min 32 seconds to add the single VM.... The VM itself was added after about 10 seconds...it spent the rest of the time doing who knows what... until it decided that the it had loaded the inventory.

Ive uploaded job logs from both the single VM and the entire group...last run took 1hr 15mins to load 55 of the VMs...
Current run is 50 VMs in 1hr 1min.. This job will fail again most likely as there isn't enough time left in the Window to actually backup.

It has gotten to the point that the load time is so slow it cant complete the job in the backup window now... We have been forced to return to DPM (of all POS things) for backups of the HyperV VMs...Shoot me please!!

PS. Everything is 10Gbe on the backend.. WIth 2x LACP links for everything

[johnnywisse]

We got exactly the same problem. Did you got already an answer from Veeam?
60 VM's with 100 CSV and just collecting all the stuff to start the backup took last night 2 hours.
When we do an inventory of the hosts in the cluster we see also that he takes a long time for locking the disks.

[veremin]

Hi, Johnny,

All environments are quite different for each other. So are causes of issues that sometimes look similarly.

Therefore, I'd recommend opening your own ticket and letting our engineers confirm your environments.

Thanks.