Active Full Backups + Wan Accelerator

[aoakley]

Hey All,

I am currently scheduled to start Active Full backup jobs once a month locally and am wondering when my next Backup Copy (W/ WAN Acceleration) job runs to offsite storage after this Active Full. Will the full amount of data from the active full (~5TB) be pushed out over the WAN or will Veeam utilize the cache on our WAN Accelerator?

Cheers

[Shestakov]

Hi,
Correct, WAN Accelerator is made to reduce amount of data sent by backupcopy job over the network. After the first run the .vbk will be transmitted, but after next active fulls only changes will be transmitted, not whole file.
Thanks!

[NightBird]

Nothing to do with wan accel or not

Backup copy job is another backup chain and do not do active full backup so only change will be push through the wan

With v9 it will be possible to do active full on backup copy job Instead of synthetic one

Boris

[BlueComp]

I want to do a backup copy over WAN, due to size/bandwidth I will need to seed the initial backup copy. I will be using forward incremental forever method, but I want to do weekly active full backups (to guard against storage corruption). Will the active fulls break the backup copy, ie. the entire full backup will have to be transferred over the WAN?

How can I minimise data over the WAN while also minimising the risks of corruption with incremental backups copied offsite?

[DGrinev]

Hi,

Active full isn't going to break the copy job as it doesn't have to be transferred over the WAN. The copy job will move only the latest state of VMs to the target.
There is a separate configuration for the copy job active fulls.
I highly recommend you to use synthetic full method for the copy job as it will reduce the amount of data transferred over the WAN.
To avoid corrupted chains there is a health-check configuration in advanced settings.

Thanks!

[BlueComp]

Thankyou, that answers my questions.

[mkretzer]

I still wonder: how does this work? How does Veeam know what changed - it cannot utilize VMware CBT can it? Even if it would use CBT it still works if VMware CBT is reset as far as i can remember... Does it scan the entire backup file on the source? That would mean alot of IO which we did not see in the past in that kind of situations...

[foggy]

Veeam B&R keeps metadata for every backup, containing block checksums, which are used to define the data changed since the last cycle.

[mkretzer]

Your product is just excellent!

That means i can have a backup which only copies the data every few months even if i do not have all the increments with the changes because it can always use the latest backup and get the changes which were done since the last copy?

[foggy]

Correct. Backup copy can have its own schedule with longer copy interval than the corresponding source backup job has. It will retrieve only changed blocks to build the restore point containing the latest VM state available in source repository. The size of this incremental restore point, though, can be comparable to the size of the full backup, due to the amount of changes occurred (if we're talking about several months copy interval).