RHEE 6.4 netwrok setup on Cisco blades

Latest response

I'm new to this env so please bare with me.

I'm installing RH on our Cisco B series blades using icsci boot from storage nodes.

On the Cisco side we created a IP address pool (10.xx.34.73 to .79),  and associated isci targets to each blade.  From there in a KVM console we mapped to RH iso and installed the OS.

In looking at the network settings, it has

- eth0 configured to 10.128.0.22 which is part of the vlan segment used for the blades to talk to the blade fabric switches.

- ibft0 config'd to 10.128.0.42

- lo  loopback.

For network setup - do I now need to configure another interface for the network on the same 34 subnet or is the server using say ...34.73 if that is what is assigned to the blade from the IP pool on the UCS side?

From my understanding the IP Pool of addresses is only used for internal comms (blades to fabric sw) and not from/to the corporate network. I was told I need to config another interface (eth1) which I've tried, giving it a ...34.60 address but the interface won't come up.  

Can anyone shed some light on this?

Thanks

Norm

Responses

It's not 100% clear from your post what your network topology is, your available physical interfaces are and what your ultimate networking goals are (throughput, redundancy, etc.).

In general, when configuring any servers to use IP-based storage, you'll want to split your application IP traffic from your storage IP traffic. Storage IP traffic should, to the greatest degree possible, be "direct" (that is, not routed). As such, you'd want to have your RHEL host configured with IP addresses in two (V)LAN segments with only the application IP address configured to route packets (the interface that's associated with your hosts's default route).