|
Problems with NIC teaming in Red Hat Enterprise 7.2
Hello --
One of our servers, a Dell PowerEdge R730xd is running the Red Hat Enterprise 7.2 distribution. The server's motherboard has for NIC ports:
em1
em2
em3
em4
As well as an expansion card with for NIC ports:
p2p1
p2p2
p2p3
p2p4
We wanted to set up a NIC team utilizing the four ports on the expansion card. To that end, four network drops were set up and connected to the four ports on the expansion card. The NIC team was set up via the nmcli commands with the LACP protocol. However, the network switch to which the server is connected, needed to have the four ports configured with the LACP protocol in order for the teaming to work. Our Network Engineering team was contacted, and I was on the phone with one of their technicians today with the intent of his configuring the switch with the LACP protocol for the ports in question.
We attempted to implement the NIC team this morning, but it failed to go into effect. The problem appears to be partly due to the network switch not being able to handle the LACP configuration of the four NIC ports correctly. The network engineering group is looking into this, and it may result in a flash upgrade.
During the course of this project there were several issues that came up on the server side:
1. After the switch was put back to its original configuration the server was still unable to connect to the network. I tried interactively bringing down and then bringing back up one of the NIC’s, p2p1, via the ifdown command. However, while the NIC was successfully brought down, it did not come back up. This process was done as the root user.
2. I initiated a warm reboot of the server via the iDrac console. The server did come back up. However, none of the expansion card ports, p2p1 through p2p4 came back up, and the NIC team that I had set up via the command line was no longer present. I was forced to disconnect one of the network cables from the expansion card, and plug it back into the em1 port on the motherboard in order to restore connectivity to the network.
I had several questions concerning this situation:
1. Why can't the individual nic's be brought back up in the manner that I described? If that is not the correct method, then what is the preferred way?
2. Why wasn't the NIC team that was set up retained after a warm boot? Is this normal? If so, is there a way to retain the configuration, or should I consider writing a script to automate the recreation of the team?
|
