Backup failure to Linux Repository

[qman_being]

Good day,

I have a strange issue backing up to a Linux based repository.

The Linux repository is an Oracle Cloud instance VM. My Veeam server is located onsite and communicates with the Linux repository over the internet (no VPN).

Adding the repository was fine however when attempting to backup to it it results in a failure below:

2021/05/12 17:43:34 :: Processing 10.4.20.15 Error: Agent: Failed to process method {MessageChannel.EstablishConnectionByRoles}: Failed to connect to the port [10.0.0.144:2501].

Now what is strange is that there is a folder created on the repo with the backup name which seems to indicate that it can communicate with the server somewhat. The other strange part of the error is that it shows the local LAN IP of the Oracle Cloud VM.

I disabled IP tables entirely on the LInux repo and allowed ports 2500-5000 inbound on the Oracle Firewall. I then eventually allowed all traffic on the Oracle Firewall inbound to the Linux VM.

I have opened a case with Veeam support - Case #04805045

[Natalia Lupacheva]

Hello,

I see you've already checked the required ports. You can also try to reach this IP via tracert, but in general I would say you should keep working with Support, because it's hard to guess what's happened.

Thank you!

[PTide]

Hi,

Have you tried specifying a preferred network?

Thanks!

[qman_being]

I have tried that option but with no luck.

I installed SMB on the Linux machine and setup the repo using the SMB share however the backup job to this still fails with the following error:

2021/05/14 18:36:19 :: Processing 10.4.20.15 Error: Agent: Failed to process method {MetaExtent.InitBackupStorage}: The network path was not found.

I cannot fathom this as I can browse to this share fine from the Veeam server itself

[soncscy]

Hey qman_being,

>I cannot fathom this as I can browse to this share fine from the Veeam server itself

I can! SMB in particular is very specific when it comes to permissions handling, and if you haven't considered every single element of the transfer process, it's easy to miss which connection is really failing.

Similarly, cloud providers (Oracle included) often have very "goofy" filter rules for public devices accessing cloud devices.

If it were me, I'd just do a Wireshark capture and see exactly what has trouble. I can only guess what the Veeam methods here mean, but phrases like "network not found", that sounds more Microsoftian, so either Veeam copied the error message, or it's just forwarding the OS error to you.

But regardless, I'm very much so a "what's actually happening" kind of guy -- Wireshark with your network team and see where the connection drops. I can almost guarantee it will be interesting. I can't see how something on the application layer can return such errors from the OS.