replicate to emergency-location

[hansis1]

Hello!

I want to replicate some Servers over WAN to an emergency-location. I have built a job an it runs well.
But the second run should be an incremental replication. But it was copied the hole file (260 GB)

What settings do I have to do?
changed block tracking size is enabled.

thx.
hansi

[Gostev]

Hansi, why do you think it has copied the whole file? It only copied as much data as the VRB file size it produced on the target ESX storage.

[hansis1]

Hello!

Thanks for reply.

The first replication needs 3 hours 43 minutes (20 MB/sec).
The second replication needs 3 hours 07 minutes (24 MB/sec).

And the Processed size was both 260,01 GB .

I think something is wrong in the settings.

thx
hansi

[Gostev]

Hansi, don't worry about the settings as you can't really control, all replication cycles are always incremental after full replication.

Don't pay attention to the processed size counter, it always reflects source data size. Incremental replication only transfers the amount of data equal to the produced VRB file size. You should look up the VRB size on target storage to determine that.

Even if changed block tracking is enabled in the settings, it only works for ESX4 hosts. For ESX3, whole VM has to be read by Veeam Backup in order to determine changes, which is why incremental backup time is pretty close to full.

Thank you!

[hansis1]

many thanks to reply.

how can I check this on target-host?
the time of first and second synchronisation is the same.

another question: what do you mean with produced VRB file size?

thx.
hansi

[Gostev]

Hansi, you just drill down to the target datastore in the Veeam Backup Servers tree, and look for the latest VRB file in the replica folder. The size of this VRB is equal to how much data was transferred from source to target during the latest incremental replication cycle.

As I've said, it is normal for the time of first and second synchronisation to be about the same if you are on ESX3.5, because whole source VM needs to be read each time to determine the changes since last synchronization. However, after the changes are determined, only changed blocks are sent over the network to the target host.

Thanks!

[hansis1]

hello!

many thanks.

i have logged the last week.
the size of the file replica.vbk is always 16.426KB.
But the .vrb-files have different sizes. These must be the versions for rollback, isn't it?

My question: The size of the .vrb-file is the size of replication?

thx
hansi

[Gostev]

Hansi, yes the size of the latest VRB is equal to the amount of data transferred during the latest incremental replication cycle. But don't pay attention to replica.vbk, it is auxiliary file that we are using to store some internal information related to specific replica. Actual disk files for replica are VMDK files.

[hansis1]

Hello!

many thanks.

But we have a problem with the Processing rate: 23 MB/s.
What can I do, that I get a higher rate?

thx.
hansi

[vbussiro]

You should definitely use the vstorage API mode with changed block tracking. There are some prerequisites, biggest are ESX(i) 4 and virtual hardware 7.
You'll see an huge speed boost (at least 10x...)