Keep high retention, but archive off older backups?

[briggs]

Hi,
Since using Veeam, we've been experimenting with the backup methods. We backup to disk, so we started off with reverse incremental. It seems like the perfect choice.
But, because of the performance deficit (takes significantly longer than traditional incrementals) and the problems we had replicating the backups to our second NAS box (the latest full backup file was seen as a new, huge file which took ages to sync), we went back to traditional incrementals with periodic active fulls (every month). We want to retain as many backups as possible (we had 6 months before with Arcserve D2D), but doing periodic fulls is impossible to maintain because of the size of these files.
I guess we could do synthetic fulls, but I was concerned experimenting with this. In my mind, an active full is pretty safe, if you have more than one, it's even safer. Synthetic fulls, say one of the files was corrupt, surely synthetic fulls can ruin a whole string of backups?

I though the obvious solution was to get Veeam to move off it's last string of backups every time it finishes the next scheduled full? We could probably cope with that, having a cheap disk array that wasn't replicated, for backups older than a month or two?

Is there anything like that in the product, because I can't find it?

Many thanks, Alan

[veremin]

(the latest full backup file was seen as a new, huge file which took ages to sync

In order to use some kind of block level synchronization tool you have to enable previous naming convention. This will guarantee that only block level changes will be transferred and not the backup file as a whole.

Previous naming convention can be enabled by the means of regkey mentioned in the following KB article:

http://cp.veeam.com/knowledgebase/article/KB1076/

Thanks.

[foggy]

I though the obvious solution was to get Veeam to move off it's last string of backups every time it finishes the next scheduled full? We could probably cope with that, having a cheap disk array that wasn't replicated, for backups older than a month or two?

Is there anything like that in the product, because I can't find it?

Currently you can have some short retention on your primary backup storage and use post-job activity or manually scheduled task to periodically offload the backup chain to another location that will store the number of restore points your need. You can even empty the storage prior to starting next backup chain with the help of the ForceCreateMissingVBK key, though it is not generally recommended.

Also it should be noted that v7 will introduce some new capabilities right in this area with its new backup copy job.

[foggy]

I guess we could do synthetic fulls, but I was concerned experimenting with this. In my mind, an active full is pretty safe, if you have more than one, it's even safer. Synthetic fulls, say one of the files was corrupt, surely synthetic fulls can ruin a whole string of backups?

Additionally, regarding your safety concern, the following thread could be useful.