TruCluster
cancel
Showing results for 
Search instead for 
Did you mean: 

ics0 interface configuration issues

SOLVED
Go to solution
jesse_45
Occasional Visitor

ics0 interface configuration issues

I have some newbie questions regarding a 2 node cluster with directly connected ics0 intefaces.

uname -a : OSF1 node1.mydomain.com.au V5.1 2650 alpha (HP TRU64 5.1b)

1) When node1 is rebooted the ics0 interface is not configured with an IP address and subnet mask whereas

node2 is. How do I get node1 to retain its network configuration over a reboot?

2) node1-ics0 is set to 10.0.0.1/255.0.0.0
node2-ics0 is set to 10.0.0.2/255.255.255.0
cluster_cnx is set to 10.0.0.42

I believe there are better values for the above. What would the best network and routing values be given

a 2 node cluster with direct memory channel connection for the ics0 interfaces on node1 and node2?

My first cut would be:
node1 : 10.0.0.1/255.255.255.0
node2 : 10.0.0.2/255.255.255.0
cluster_cnx : 10.0.0.42

"route add -net 10.0.0.0/24 -dev ics0" configured somewhere????

All other addresses 10.0.1.0 and above to route out through the default gateway interface (nr0).

The other interfaces (nr0,nr1) use a class C private network.

What other gotchas do I need to consider when configuring these interfaces?
6 REPLIES
Ivan Ferreira
Honored Contributor

Re: ics0 interface configuration issues

In a TruCluster environment, the ics0 interfaces normally uses 255.255.255.0 network mask. I don't know why node1-ics0 is set 255.0.0.0. I think is incorrect.

I don't know the function of cluster_cnx. Is not normal to have this host in the ics network.

You don't need "route add -net 10.0.0.0/24 -dev ics0", these are implicit routes added when the interface is configured. You can check your routes configuration in /etc/routes.

Why your default gateway is your netrain interface?
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
Mark Poeschl_2
Honored Contributor

Re: ics0 interface configuration issues

10.0.0.1/255.255.255.0 is absolutely the right answer for the configuration of node1's cluster interconnect.

I don't think the 'cluster_cnx' address is relevant - unless of course you have a CAA service or application that uses it. It's not part of the default TruCluster setup.

You should not need to manually add any routes for the 10.0.0 network - that should get taken care of automagically once the ics0 devices are configured properly on both nodes.
jesse_45
Occasional Visitor

Re: ics0 interface configuration issues

Thanks you Ivan and Mark for confirming the node1-ics0 interface parameters.

Is anyone able to tell me why the node1-ics0 interface is not being configureed at boot time and how to get the node1-ics0 interface configuration to survive a reboot? Or, where I should look to troubleshoot the problem?

Yes, the default gateway is on the netrain interface nr0. The nr0 interface configuration does survive a reboot.
Ivan Ferreira
Honored Contributor
Solution

Re: ics0 interface configuration issues

In the /etc/rc.config file, you must have the following entries:

NETDEV_0="ics0"
export NETDEV_0
IFCONFIG_0="10.0.0.1 netmask 255.255.255.0 filter"
export IFCONFIG_0
Por que hacerlo dificil si es posible hacerlo facil? - Why do it the hard way, when you can do it the easy way?
jesse_45
Occasional Visitor

Re: ics0 interface configuration issues

Many thanks for pointing me in the right direction.

/etc/rc.config was a bit of a mess. Sorted that out using rcmgr and all is well.
jesse_45
Occasional Visitor

Re: ics0 interface configuration issues

sorted out /etc/rc.config using rcmgr