short info at the start, English is not my mother language ;).
I have a problem with 2 Supermicro Server. Both are connected to a DELL S4148F-ON Switch with Force 10 on it. Both Servers are Ubuntu
20.04.04 LTS, both have the same configuration.
They have also an Beegfs configured and an Infiniband connection to an HPC Cluster. My Problem is, that both Server unable to start the network adapter/connection. I did a lot of troubleshooting, i tried to paste the journactl output here, but when i try to post this question i got the information that this question looks like spam. How can i provide here the output?
Our target adapter is enp68s0f2, that is the network where the communication is working for the beegfs, later we want to change it to the Infiniband connection.
The Problem is that network is not working as it should, also all services like beegfs if there is no network.
What i found out is, when i change the vlan from 30(the correct vlan for this interface) to an other vlan , the network adapter is start working fine. When i change then back to the correct vlan , the server are able to ping and network is working as it should. But that is to late for the beegfs and multipath stuff. It makes no sense to switch the vlan (untagged) to an other and the adapter is start working fine. If i shutdown this port on the switch and enable the port, is there no change and the adapter is not working. To restart the network service is also not working. I have to change the vlan, or (sorry i forgot to mention) if i restart one of the sever the other one is start to working.
Hopefully, it is understandable what my problem ist :). That is a really strange behavior.
Thank you for any suggestions and help :).
SG