Slow VM replication even with local seed

[itlad]

Hi all,

I'm trying to replicate a VM from a remote site to main site,
but using a seed backup which was taken 2 weeks ago, however it's running very very slow.

I have selected replica seeding option,
specified the local NAS repository on the main site as the source of the seed,
specified the remote site server as the repository for the replica metadata,
and also specified source and target proxies as local resources to each site.

The restore of the replica from the seed runs fine, the 'calculating digests' phase runs reasonably fast,
but the next phase runs at only 157kb meaning it would take 2 whole days to complete a single VM with 30GB disk.
(I have to migrate about 2TB in total)

Any ideas why so slow?

[tsightler]

The screenshot shows that the bottleneck is the "Network". What is the speed of the WAN link between the sites? The screenshot implies that it might be something slow like a T1.

[itlad]

The screenshot shows that the bottleneck is the "Network". What is the speed of the WAN link between the sites? The screenshot implies that it might be something slow like a T1.

it's a 10mb ADSL connection with 3MB upload bottleneck.
150kb is only 1 twentieth of 3mb so I'm not convinced the WAN bottleneck is to blame.

[dellock6]

Are you sure about your numbers? Usually network connections are misured in bits, while Veeam jobs shows bytes.
A 3 mbits connection equals to 384 KBs, this is the fastest speed you can get from this line. 157 is half of the maximum (supposed) speed, if there is other traffic on the wan link the speed is quite correct.

[itlad]

Are you sure about your numbers? Usually network connections are misured in bits, while Veeam jobs shows bytes.
A 3 mbits connection equals to 384 KBs, this is the fastest speed you can get from this line. 157 is half of the maximum (supposed) speed, if there is other traffic on the wan link the speed is quite correct.

dellock6 you are right of course - I am confusing myself
We have a higher bandwidth fibre connection which I plan to switch over shortly so that bottleneck should "widen".

In any case, what I'd like to understand now is what exactly is it transferring over the WAN when an incremental backup is run?
This was only a test and since the initial seed backup was taken, only a desktop shortcut and single registry key was added to validate that the incremental works.

[tsightler]

Remember that Veeam transfers "blocks", not files. Blocks are changing on disks all the time due to logging activity, reboots, attribute updates, etc. You will probably want to let the transfer complete as, if there are truly only a small amount of changes on a small VM there will be periods of slow performance while changed data is being transferred, and then periods of fast performance when blocks are the same. Did you select "WAN target" for your storage optimizations and "Best" compression?

Also, make 100% sure that you selected the correct proxies (it appears from the screenshot that this is correct, but it's by far the most common error seen in the field).

[itlad]

Remember that Veeam transfers "blocks", not files. Blocks are changing on disks all the time due to logging activity, reboots, attribute updates, etc. You will probably want to let the transfer complete as, if there are truly only a small amount of changes on a small VM there will be periods of slow performance while changed data is being transferred, and then periods of fast performance when blocks are the same. Did you select "WAN target" for your storage optimizations and "Best" compression?

Also, make 100% sure that you selected the correct proxies (it appears from the screenshot that this is correct, but it's by far the most common error seen in the field).

Yes I'm sure correct proxies were specified- found that it was even worse when leaving proxies as Auto
I didn't change compression & storage optimization settings but will do so now...

Thanks