Configuring a standby network interface fails

[Steven Sameshima]

I'm having problems configuring a standby network interface for servers running RH AS 3 update 6 and MC/SG A11.15.05. Note that I've done this many times with HP-UX.

In my cluster configuration file I have something like the following:

NODE_NAME abcla1a
NETWORK_INTERFACE eth0
NETWORK_INTERFACE eth2
HEARTBEAT_IP 192.1.1.1
NETWORK_INTERFACE eth1
NETWORK_INTERFACE eth3
HEARTBEAT_IP 193.1.1.1

NODE_NAME abcla2a
NETWORK_INTERFACE eth0
NETWORK_INTERFACE eth2
HEARTBEAT_IP 192.1.1.2
NETWORK_INTERFACE eth1
NETWORK_INTERFACE eth3
HEARTBEAT_IP 193.1.1.2

When I try and build the cluster configuration file I get the following output:

Checking cluster file: /usr/local/cmcluster/conf/abcla1_cluster.config
Checking nodes ... Done
Checking existing configuration ... Done
Warning: Can not find configuration for cluster abcla1_cluster
Gathering configuration information ... Done
Gathering configuration information ..
Gathering storage information ..
.
Gathering Network Configuration ........... Done
Error: Interface eth0 on abcla1a has an IP address on it
(192.1.1.1), but the configuration file doesn't have it.
This may be caused by a local switch or changes in the network configuration.
Error: Network interface eth2 cannot be found on node abcla1a
Error: Interface eth0 on abcla2a has an IP address on it
(192.1.1.2), but the configuration file doesn't have it.
This may be caused by a local switch or changes in the network configuration.
Error: Network interface eth2 cannot be found on node abcla2a
Error: eth0 on node abcla1a cannot be configured in the cluster
because it does not have an IP address, and it is not a standby lan for any other lan.
Error: eth1 on node abcla1a cannot be configured in the cluster
because it does not have an IP address, and it is not a standby lan for any other lan.
Error: eth0 on node abcla2a cannot be configured in the cluster
because it does not have an IP address, and it is not a standby lan for any other lan.
Error: eth1 on node abcla2a cannot be configured in the cluster
because it does not have an IP address, and it is not a standby lan for any other lan.
cmapplyconf : Unable to reconcile configuration file /usr/local/cmcluster/conf/abcla1_cluster.config
with discovered configuration information.

Warning: Network interface eth0 on node abcla1a couldn't talk to itself.
Warning: Network interface eth1 on node abcla1a couldn't talk to itself.
Warning: Network interface eth0 on node abcla2a couldn't talk to itself.
Warning: Network interface eth1 on node abcla2a couldn't talk to itself.
Failed to evaluate network

Does anybody know what I'm doing wrong?

[melvyn burnard · Accepted solution]

SG on Linux does not do the network side the same way as HP-UX.
You need to use the Linux functionality known as bonding, and create a bond using a "primary" and a "standby" Lan, then set up the bond as the network in the SG configuration.
This is all documented in the Managing Serviceguard Manual on docs.hp.com/en/ha, making sure you read the manual for Linux, not HP-UX

[Steven Sameshima]

This is what I get for not reading the documentation. Actually the last and only time I read the doc for the Linux version of MC/SG was for A11.14 on RH 7.3. I'm familiar with IP bonding so I'll go ahead and use that. Thanks for pointing me in the right direction.

[Steven Sameshima]

Use IP bonding