Problem while creating a SNC Cluster

[RAJESH GANGADHARAN]

Hi All,

We are trying to build a Single Node Cluster. But while I am giving the cmcheckconf command, it is giving following error.

Begin cluster verification...
Note : a NODE_TIMEOUT value of 2000000 was found in line 117. For a
significant portion of installations, a higher setting is more appropriate.
Refer to the comments in the cluster configuration ascii file or ServiceGuard
manual for more information on this parameter.
Error: Unable to establish communication to node XXXXXX: 19 (where XXXXXX is my node name )
cmcheckconf : Failed to gather configuration information

I can avoid NODE_TIMEOUT value, thats is not a proble. I am using DNS , and when I am changing the same to local files ( by editting /etc/nsswitch.conf ) I could pass this test and I can create a cluster. There is no problem with the entris in the /etc/hosts, /etc/cmcluster/cmclnodelist, and my cluster configuration file, the reason why I am teling this is I could create a cluster without DNS.

When I give cmquerycl ( which will not check any configuration file), I am getting the following error message.

root:>cmquerycl -v -k -n XXXXXX ( Node name )

Begin checking the nodes...
Error: Permission denied to 127.0.0.1
Error: Unable to establish communication to node XXXXXX: 19
Failed to gather configuration information.

Any help will be greatly appreciated.

Thanks in Advance...........

[Uday_S_Ankolekar]

This is probably due to hacl entry in /var/adm/inetd.sec file. In that case insert 127.0.0.1 infront of hacl entry
and then try cmquerycl. You can also try cmscancl for detailed information.

-USA..

[RAJESH GANGADHARAN]

Uday,

I don't have /var/adm/inetd.sec file. I hope in that case it will allow for all. Correct me if I am wrong

[melvyn burnard]

Which version of Serviceguard are you running?
If it is 11.16, then you need to be aware that you cannot use ip addresses in your cmclnodelist file unless you have installed the latest patch.
Also be aware you mmay need to have the relevant hostname iserted in the file for every configured network on your node.

Please take a read off:
http://docs.hp.com/en/6283/SGsecurityfiles.pdf
http://docs.hp.com/en/5874/securingserviceguard.pdf
http://docs.hp.com/en/B3935-90078/B3935-90078.pdf

[RAJESH GANGADHARAN]

Hi melvyn,

My SG version is 11.15.

I have already populated the host name details in the /etc/hosts file. As I have mentioned, this problem witll occure when I am using dns ( in /etc/nsswitch.conf ) for host name resolution. If I put files ( for local /etc/hosts ) or host name resolution all are working perfectly file.

Here I am trying to create a Single node Cluster. I got multiple SNCs and two metro clusters in my setup. Every where we are using DNS oly for the name resolution.

Any catch.....

[melvyn burnard]

If it works fine without DNS then you have a DNS issue.
The general recommendation for any Serviceguard cluster is to use /etc/hosts first, and then DNS.
This prevents cluster issues if DNS becomes unavailable, invalid etc.

[RAJESH GANGADHARAN]

Hi Melvyn,

You are right in one way. But while I am giving a wild card in the cmclnodelist, it works perfectly fine. But I got 8-10 similar setup of SNC cluster which works perfectly fine with same configuration. It seems to me the only difference between this machine and other machine is a patch, PHSS_31067. Currently I am working with HP for getting some clear solution for my problem.

Thanks..................

[RAJESH GANGADHARAN]

Hi All,

The problem was with one of the patch PHSS_31067.I have compared this with some of my other server. Now machine is working perfectly fine. Thanks very much for all who worked on this thread.


Cheeers............