Endpoint to B&R repository fails.

[d.lansinklesscher]

I have two networks that are not connected/routed:
10.254.0.0/24
172.16.16.0/24

The Situation:

I'm running B&R V8 patch 2 and endpoint. These are my compononents:

Name: BK-SVR03.homed.local
Role: Windows B&R server
Nic1: 10.254.0.105
Nic2: 172.16.16.51

Name: BHR-BK01.beheer.local
Role: Linux Datamover
Nic1: 10.254.0.159
Nic2: 172.16.16.52

Name: BHR-DB-01.beheer.local
Role: Machine that will be backed up with Endpoint.
Nic1 : 172.16.16.21

The Problem
I configured an endpoint job on BHR-DB-01. This job uses a B&R repository on BK-SVR03 which is added by its IP 172.16.16.51. The repository resides on BHR-BK-01.beheer.local. So far so good.

When I launch the job it fails with the error:
A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond 10.254.0.159:2502

This error makes sense because the IP is ofcourse not reachable on BHR-DB-01. What I don't get is why it connects to this IP because BHR-BK01.beheer.local resolves to 172.16.16.52 on the BHR-DB-01. Is this somehow fixable?

[JaxIsland7575]

I have a similar setup with two networks connected to my servers, I have had to use persistent routes or hosts file entries on some clients to force them to use a specific IP. Your setup is not identical to mine but it may help.

[skochetkov]

Hello from Veeam Endpoint support.

The backup job failed due timeout, so lets try this workaround: run regedit on the Veeam EndPoint side, go to HKEY_LOCAL_MACHINE\SOFTWARE\Veeam\Veeam Endpoint Backup and create DWOD value ConnectByIPsTimeoutSec=1000, restart Veeam EndPoint service and try once again

[d.lansinklesscher]

Hi Sergey,

Sorry for the delayed response. I had already changed my networking to get it working. However today I'm trying to do a baremetal restore , and in the recovery I have only one subnet available. While trying the restore I run in to the same issue. Enpoint tries to connect to an IP in a range that is not connected to the physical server I'm running baremetal restore on.
I have added the registry key in recovery but it makes no difference.

[Dima P.]

Hello Dennis,

Please check via Tools > CMD after the Network is configured in the restore wizard (you can simply close the wizard and go to tools) if the backup repository is "pingable". Thanks.

[ninjaburn81]

Dima:

Having the exact same issue, and this worked for me. Thanks!

[Dima P.]

Adam,
You mean adding the registry key while booted from the Recovery Media?

[d.lansinklesscher]

The repository is pingable, on its ip in 192.168.69.x /24 because the recovery image has an IP in that subnet too. However the repository also has another address in another subnet 10.254.0.x /24. This IP is not pingable because the recovery image is not connected to this network. The problem is that endpoint in the recovery image tries to connect to the IP in the 10.254.0.x /24 subnet and this of course doesn't work.

And yes I added the registry key while I was booted from recovery image.

[ninjaburn81]

Dima:

Yes, I added the registry key, and it worked to both a Windows share and an SMB (Data Domain target) share. (I was getting failures before on both.) One thing of note: Not sure why it is taking so long to run ...it took 50 minutes to do a simple file backup (first seed) of about 40 MB for both targets. Client, server and targets are all on the same subnet, 1GB lines, not heavy amount of traffic. . A incremental takes about 2-3 minutes for the same job, so a bit concerned how long a full backup will take. I will test that and update.

What is the registry key doing exactly?

[Dima P.]

d.lansinklesscher,

recovery image tries to connect to the IP in the 10.254.0.x /24 subnet and this of course doesn't work.

What happens if you manually configure the adapter at the Network step of the wizrard to use the proper subnet and IP address?

[Dima P.]

Adam,
It does what it says - increase the timeout value while getting a response from the host

it took 50 minutes to do a simple file backup (first seed) of about 40 MB for both targets

Same performance for VBR Repository and Shared folder destinations?