reason for not backing up / replicating RAM

[mcz]

Hello everybody,

I'm wondering why the veeam-guy's took the decision for not backing up the RAM of the vm to be able to restore it later and resume exactly at the point where the snapshot has been taken... Consider applications/databases which are not VSS aware, currently you always have to stop these services and start them after snapshot has been taken to make sure they are consistent. If you would backup RAM and restore it later, you would be able to continue your OS and the application was fine.

Next consideration is that sometimes it takes a lot of time to get all the services of a vm started and available, in case of RAM restore everything would run already...

Thanks for your answers!

[Andreas Neufert]

Hi Michael,

thank you for the question.
If I remember right the VM will be stunned until the memory is written to the disk, so your database would not be accessible during that time.
https://kb.vmware.com/selfservice/micro ... Id=1015180

As well there is the best practice to let the OS and Application know that there was a restore, so that they can replicate with other databases or Application metadata in a way without corruption.

Overall nearly all databases can be brought in a consistency state with a command for the time of a VM Snapshot (some seconds). This should be the optimal way. If you tell us which database you have, we maybe can help with a script or Veeam configuration to achieve this.

[mcz]

Hi Andreas,

thanks for the clarifications. Well in our case (we use vsphere 6.0) the impact to the vm/database is less than a second during which the vm won't respond... We use an old SqlBase database version 9.0.1 which has no vss support...