Re: Clufvy and private interconnects

From: Dan Norris <dannorris_at_dannorris.com>
Date: Wed, 12 Mar 2008 07:16:34 -0500
Message-ID: <47D7C9A2.8050900@dannorris.com>




  


I'm not sure why it is failing. In previous versions, when you used RFC
reserved networks, the tool couldn't find a network suitable for VIPs
since it assumed that all reserved networks were for private
interconnects. Looks like you sort of have the opposite problem. I'm
not sure the reason for this and there's no other notes on ML
indicating why, so I'd file an SR on it. I expect that VIPCA should
still run fine anyway. 

I noticed that you have 2 NICs configured with separate IPs on each of the two networks. If you want to build some redundancy for the NICs on your servers, you need to investigate interface bonding. Simply putting two NICs on the same subnet with different IP addresses isn't sufficient to create a redundant NIC configuration. Instead, you'll need to bond the two physical interfaces together (with the bonding software driver) and then use a single IP address on the bonded pseudointerface (typically called bond0, bond1, etc). Search metalink for "linux ethernet bonding" and you'll find a few helpful notes.

Dan

Jeffery Thomas wrote:
Solaris 10  We are in the process of prepping two boxes for a 10g RAC cluster.   I downloaded the 11g cluvfy (as recommended by Oracle) and our config passes every check; but for some reason cannot find suitable interconnects.   

My question would be:  what exactly is cluvfy looking for when it is scanning for the interconnects?    User equivalence checks out, we are using switches, and so on.

cluvfy comp nodecon -n oracnp01,oracnp02  -verbose

<public IP stuff found)
....
Interfaces found on subnet "192.168.1.0" that are likely candidates for VIP:
oracnp01 e1000g2:192.168.1.101
oracnp02 e1000g2:192.168.1.103

Interfaces found on subnet "192.168.1.0" that are likely candidates for VIP:
oracnp01 e1000g5:192.168.1.100
oracnp02 e1000g5:192.168.1.102

Interfaces found on subnet "192.168.2.0" that are likely candidates for VIP:
oracnp01 e1000g3:192.168.2.101
oracnp02 e1000g3:192.168.2.103

Interfaces found on subnet "192.168.2.0" that are likely candidates for VIP:
oracnp01 e1000g4:192.168.2.100
oracnp02 e1000g4:192.168.2.102

WARNING: Could not find a suitable set of interfaces for the private interconnect.

Result: Node connectivity check passed.

Thanks,
Jeff
-- http://www.freelists.org/webpage/oracle-l Received on Wed Mar 12 2008 - 07:16:34 CDT

Original text of this message