Server not reachable: networking issue

[Helmonder]

Hi !

 

After a reboot my server was not reachable over the network anymore.

 

Server is reachable over IPMI and I can use the local console

The server itself is running fine I can just not reach it.

 

The issue arose when I activated a 10GB Mellanox card for my two urnaids to talk to each other. After the reboot I guess the server thinks that the 10GB mellanox connection should be primary because it is faster. I did set a higher metric on it though.

 

I have copied back old network.cfg and network-rules.cfg files from a backup, that seems to have worked. Do not really get it though, They have not been changed since the backup (cards were allready in)..

 

I have added diagnostics.. Can someone help me in checking what is wrong with my networking config ?

 

My setup is as follows:

 

ETH0/BR0 is a bridged interface consisting of eth0 (which is a 10GB Mellanox to my network). At some pointed I wanted a cabled 1GB in eth1 as backup in this bridge, that interface is disabled, do not really remember what I did there. This interface is 192.168.1.5

 

ETH1  is a regular 1GB network port directly on my mainboard of the server, it is inactive.

 

ETH2 is a 10GB Mellanox direct connection to my other server, it is 192.168.10.5 and only talks to 192.168.10.6. This connection is only used for the two servers to talk to each other and this works fine.

 

ETH3 is the second interface of the Mellanox card, empty and not used.

 

 

 

 

 

 

 

 

 

 

 

 

 

tower-diagnostics-20210116-1125.zip

Edited January 16, 2021 by Helmonder

[JorgeB]

Boot using GUI mode and reconfig LAN.

[Helmonder]

21 minutes ago, JorgeB said:

Boot using GUI mode and reconfig LAN.

Did that... GUI did not load though... I ended up being able to get it loaded by changing the routes on console though... So I am up and running but my setup is still somehow flawed, this will happen again on a reboot and I want to avoid that..

 

[JorgeB]

If you're up and running you can just reconfigure the LAN as needed, gigabit NIC should be eth0, and 10GbE configured using a different subnet and with just the IP address.

[Helmonder]

1 hour ago, JorgeB said:

If you're up and running you can just reconfigure the LAN as needed, gigabit NIC should be eth0, and 10GbE configured using a different subnet and with just the IP address.

That is what I have... still it borks when I reboot...

 

 

[JorgeB]

It shouldn't, does anything change by itself in the LAN settings page after a reboot?

[Helmonder]

Nope... looks like when doing a softreboot it works, but after a hard reboot it does not..

[Helmonder]

 

This is where the mistake is... the default route is set as 192.168.10.6 via eth2.

 

That is my internal and not externally connected network.

 

The default route should be 192.168.1.250 (my router) via eth0 (my uplink)

 

I added the correct route, but now I have two.. pressing the wastebin icon on the wrong one does not work..

 

I removed the extra route via console:

 

route del default gw 192.168.10.6 eth2

 

Its gone in settings now... But this is the way it was earlier... I -think- this will reset itself after a reboot again.. Which is also logical since this was done in the non persistent memory.. Where does unraid store its route information in boot/config ?

Edited January 16, 2021 by Helmonder

[JorgeB]

Not really a network guy, someone else might be able to help.