I have some open cases with Veeam support I'm working through, but I thought I'd throw out some of my frustrations and suggestions to find out if it's just me or if others are experiencing similar problems.
1) Copy job will fail midway through a machine if the copy time has elapsed: Would be nice if it could finish up what it was busy before doing so rather than throwing away GBs of file transfer. For example I've seen it fail at 75% processing through our exchange server, which can be a good few GBs. If you've only got 5Mbps fibre that's a pretty big setback.
2) Copy job processing order: Presently if the job stops midway and restarts it starts at the very beginning. Thus I end up with many restore points for the first few machines and progressively less for the remaining machines while the backup catches up. Would be nice if it could start where it last left off or pick the machine with the least restore points.
3) Transform takes a long time: I have grown accustomed to synthetic full which appears to run much faster than transform. Transform really hits my bulk Synology SATA storage hard. I'd rather take the knock on an extra synthetic full file on disk space than the time it takes to run the transform. I realise that due to how the copy job process data this may not be possible, but I have to ask
4) Transform failures when connectivity is interrupted: When the copy is done and the transform is running on the remote end a connectivity interruption on the client side will cause the transform to fail. It would be nice for this to be able to process without active connection from the client side. i.e. when the client reconnects the target reports back on the current status of the transform.
5) File Locks when when connectivity is interrupted: If you're going to fail the transform due to a connectivity interruption then there needs to be some time-out to kill the VeeamAgent.exe process locking the file it was using in the transform. This locked file prevents the copy job from processing any new data.
6) Copy data before transform: Often during a catchup phase with restore points I get even more behind with my copy job because a transform needs to complete first. There should be some intelligence which should prioritise getting current data accross over transforming. i.e. copy the data first and then run the transform.
In essence if ever you get behind with you copy job, it will either be interrupted by the copy interval time-out which means it starts from the beginning which could also mean getting delayed even more while a transform completes. In my opinion if you have a small link and/or slow target storage copy jobs at this stage, are not for you.
