PB with my cluster configuration file....

[Moreau Stéphane]

I have a cluster with 2 nodes (HP D330) and HP UX 10.20 / MC Service Guard A10.12 (ok, it's quite old :)

We recently changed the IP adresses, of the 2 nodes (i changed IP and SUBNET on all package files, modify /etc/hosts and netconf file etc...)

both node restarted correctly (network / LAN configuration is good), but i don't know why, i have a pb with my primary node (only on this node... i started the cluster correctly on the 2nd node : cmruncl -f -n hostname_2nd_node).

All the "cmcl" commands have the same error :

# cmviewcl
cmviewcl : Cannot read the cluster configuration file. Either it does not exist, or it is corrupted, or it is empty.

I can certify that the configuration file (cmclconfig or cmclconfig.ascii) is right because the cmapplyconf runs correctly on my secondary node, and copy the cmclconfig file on the primary node.

I checked the bin and file rights, but they are both good...

what do you think about my problem ?

Thx for your reply... (my last chance... my ignite backup)

[Ashwani Kashyap]

Can you do a cmapplyconf on the problematic node with the new IP in the ascii files .

[Moreau Stéphane]

thx for your answer, but i can't use any "cm" command on my 1st node... (cmruncl, cmviewcl, cmapplyconf, cmgetconf... don't works)

i have permission denied !! (but the right are goods...)

# cmapplyconf -v -C /etc/cmcluster/cmclconfig.ascii

Checking cluster file: /etc/cmcluster/cmclconfig.ascii
Checking nodes ... Done
Permission denied to hostname_node_1
Checking existing configuration ... Done
Permission denied to hostname_node_1
Gathering configuration information ........ Done

Error: Unable to communicate with node hostname_node_1

Error: Unable to determine device configuration: failed to send device query to node hostname_node_1

cmapplyconf : Unable to reconcile configuration file /etc/cmcluster/cmclconfig.ascii
with discovered configuration information.

:/ ??? :)

[Moreau Stéphane]

in fact i started the beside command :

cmapplyconf -v -f -C /etc/cmcluster/cmclconfig.ascii -p /etc/cmcluster/all_packages_conf

"all_package_conf" is a file wich contain the directory of the "packages" configuration file

[Ashwani Kashyap · Accepted solution]

Looks like it is not able to resolve node2 .
Check whether host files and DNS entries are proper .

I would also see a amgetcong on the working node and see what information about node1 and node1's devices I get and then check ifactually they are there .

[Moreau Stéphane]

I solved my problem...

On the 2nd node, all CM commands works perfectly... the node started correctly (cmruncl -f -n 2nd_node_hostname).

So i decided to make this :

rm -r /etc/cmcluster (on my 1st node)
tar cvf /tmp/cmcluster.tar /etc/cmcluster (on my 2nd node)
rcp the tar from my 2nd node to my 1st node
and finaly "tar xvf cmcluster.tar"

And now it's works fine on my 1st node !

perhaps one, of my configuration file, (or package file) was corrputed... I don't know.

Thx for your help :p

[melvyn burnard]

I suspect you either had no cmclnodelist, or it was missing an entry for itself, on the failing system.
By copying everything from the working box, you have replaced the cmclnodelist file with a good working one.
Do you have a cmclnodelist in your /etc/cmcluster directory?

[Moreau Stéphane]

In fact we just moved the server, and change the IP @dresses. so, normally the cmclnodelist file did not change...

But, you are right... the cmclnode file of my old "/etc/cmcluster/" directory is not good...(i check it on backup)

There is an old hostname for my primary node(we changed the hostname of the server... 2 years previously !!!) So, how my server could function correctly since 2 years, if the file was wrong !!?? i can't explain that !