Rapid Recovery

Agentless VM KEEPS doing Base image

We have about 20'ish VM running on a 3 node Hyper-V cluster, I'm using agentless backups for most of them.  I've had a few issues with a small number of the VM's but I have one with an issue I can't seem to solve.

One of the VMs KEEPS DOING BASE IMAGES! It'll do the odd incremental but always seems to do a Base.  Its just a server 2012r2 VM running our building control software (HVAC, etc) nothing special its only around 25Gb in size, and has VERY little data churn!


I could put up with the constant base images as performance, bandwidth, repo size isn't an issue, BUT as it keeps doing a base it breaks replication to our offsite Core giving a duplicate base image has been found error and replication stops.


Any ideas WHY its doing a base and how to stop it!




  • Is this happening when the VM is being migrated over to another host? If that was the case, or if you wanted to check you could migrate it over to another host to validate if that was the case (of course if it was that would create another base).

    Does it fail a backup and then take a base?

    Just curious is the hard disk a .vhd or a .vhdx file (I have seen this happen on a .vhd however that is due to the fact that it isn't supported).

    Anything else scheduled to run on the host during this VMs backup? Or is simply in the same schedule as the others?
  • nope its sitting on the same node of our cluster. The backups don't fail it just keeps doing a base, then when it tries to replicate I get a duplicate image error and the replication fails. I've deleted it again and its done a base followed by an incremental BUT this has happened before then it starts to do bases again! Currently I have it backing up once a day and nothing else is running at that time (I've spread the backups out a bit)
  • Hi Kerry:
    Probably the quickest fix is to install the agent on that machine. (Make sure that you change the display name of the protected server before removing it if you want to keep the current recovery points; this will avoid a name conflict when protecting the server again; you cannot (easily) change the name of a recovery points only instance).
    The issue is probably related to VSS issues on the Hyper-V host (duh!), specifically, the host being unable to 'link' with the VSS inside the protected machine to properly do the quescing. If you want to use agentless protection, you should open a ticket with us and we will figure it out together.
  • this eventually sorted itself out! But it took a while of removing the server and then backing it up again. Its managed to go a week now doing incrementals