1:1 Data Replication

[Baumeister]

Hi,

for last few weeks I have been building a new backup infrastructure for our company. It all went more or less well, but I have one last problem, which I have failed to solve. We have a server for data sharing, with an external RAID attached to the Hyper-V Host and a virtual hard disk mounted to the data share server VM. The VM is stored on the internal SSD drive of the host server, while the data is stored on the VHD on the big external HDD RAID. My problem now is that I cannot find a solution for backing up the data on the RAID. The backup host looks exactly the same as the source host and my goal is to have two identical machines in the end, with everything on the source machine copied exactly as it is to the target machine. But as told before I couldn't find a way of backing up the data on the RAID as I want to, i just need a exact copy of the data on the original RAID on the target RAID, with no VEEAM interpreter needed to read the data, as the plan is that it can be attached directly to the backuped server. I can't use the Copy Job because it will always read and write the complete data and we want to run the backup every 3-4 hours and we have about 30 TB data, far too much to be copied every 3-4 hours as a full Backup. So the backup has to be incremental.

Does anyone have a solution for this?

My Support Case ID is 03495542.

Thanks for your help

Andreas Baumeister

[PTide]

Hi,

Why not to use a replication job for that VM?

<...>and a virtual hard disk mounted to the data share server VM

How is it mounted? Is it added as another virtual disk in VM settings?

Thanks

[jmmarton]

From a backup perspective, use forward incrementals with synthetic fulls or forever forward incrementals. For the backup copy job, keep in mind that's always a forever forward incremental. Unless you perform an active full in the original backup job, the BCJ will only copy changes. Backing up multiple times a day doesn't change this either so there shouldn't be anything within the BCJ that's copying 30 TB of data repeatedly. Unless the 30 TB of data has a 100% change rate which I assume it doesn't.

Joe

[bally@biosparq.nl]

I am just starting to work with VM's and Veeam and backing up a VM of only 3.7 TB, much less then the 30TB. As backup processes take quite some time(full offsite backup almost one week) I have the the feeling it is difficult to get all processes running smoothly all the time. Almost all the data is measurement data, which only grows but never changes. Wouldn't it be better to have this data outside the VM, and use a simple 1:1 copy process like Robocopy or Rsync? As simple processes are more stable.

[PTide]

Hi,

As backup processes take quite some time(full offsite backup almost one week)

If you provide us with some bottleneck stats from the job then I might be bale to recommend some improvements for your workflow. Also, what is the target repository type that is offsite?

Almost all the data is measurement data, which only grows but never changes.

If it grows then it certainly changes block on the virtual disk, therefore CBT (changed block tracking) will catch those changes and transfer them as incremental restore point. That is, after full backup is done, all subsequent backups will go much faster.

Thanks!