Restore path to DMZ on another site?

[pkelly_sts]

I'm currently doing a small windows file restore to a VM in a DMZ on another site but getting shocking performance so wanted to understand the data paths.

It's literally 41Mb, 155 folders, 262 files & so far it has taken 34 mins but the last 1Mb is taking forever (claiming 26KB/s in the restore status dialog).

Full info:
Site A:
VBR server (VM)

Site B:
Local Physical Repository Server hosting this backup (also configured as mount server for all repos on Site B)
DMZ with the VM in it

VPN:
My Laptop (running VBR Console) connecting to VPN endpoint at Site A

I notice that the drive has been mounted on my laptop (c:\VeeamFLR has data visible) but in the logs I can see it mounting on my machine, then mounting on the repository server?

I know that ordinarily my laptop won't have SMB access to this DMZ server and there's a log entry stating "Restoring via vSphere guest interaction API".

I'm connected via 20Mb VPN (to Site A) and Site-A to Site-B is a 100Mb P2P link.

The thing that's confusing me is that I'd expect it to be an all-or-nothing thing, it's the first time I've tried restoring to a DMZ-based VM so wasn't sure if I'd need to open any additional ports but clearly it's connected (as I can see the restore progresisng & files appearing in teh guest, but why so god-awfully slow, and getting slower all the time (restored around another 100kb in the few mins I've been writing this post).

Sounds like it might be worth me opening a case for checking but wanted to understand the data path anyway?

[DGrinev]

Hi Paul,

Is it FLR to original or new location?
Also, do you have Mount Server associated with the repository? Thanks!

[pkelly_sts]

Yes, FLR to original location, mount server for the repository is the repository server itself, on same physical site (site B) as the DMZ server, but obviously on an internal LAN (i.e. not in the DMZ).

[DGrinev]

Your scenario and setup looks correct to keep the VM traffic in one site, but seems the data path wasn't routed as expected.
It's worth to let the support team take a closer look on the logs and reveal the root cause of the performance.

[pkelly_sts]

OK, happy to open a case, but what job do I reference (to extract the log files when running the )? I've only ever captured logs for a backup/replication job, never a restore task before?

[DGrinev]

You can just raise a case and ask support team to guide you through the process. Thanks!

[pkelly_sts]

OK will do, thanks