cmquerycl problem ...

[buell, stefan]

I am trying to form a two node cluster (SLES 8.0, MCSG 11.15)with LockLun. The node names are "node1" and "node2"
Both the nodes have .rhosts and cmclnodelist entries as follows,
------------------------------
node1.domain.de root
node2.domain.de root
------------------------------

node1:/home/admin # cmscancl -n node1 -n node2

The nodes to be scanned are: node1 node2
The output file is: /tmp/scancl.out
Checking rsh access to all the nodes...
Getting information from node1...
Getting information from node2...
Checking remote network connections (node1 to node2)...
Done. Check log file /tmp/scancl.out for all information.

Unitedlinux syntax for syslog = tail -f /var/log/messages

Feb 18 12:27:10 node1 in.rshd[4008]: root@node2.moinmoin.de as root: cmd='rcp -t /tmp/'
Feb 18 12:27:10 node1 in.rshd[4015]: connect from 212.202.154.172 (212.202.154.172)
Feb 18 12:27:10 node1 pam_rhosts_auth[4015]: allowed to root@node2.moinmoin.de as root
Feb 18 12:27:10 node1 in.rshd[4015]: root@node2.moinmoin.de as root: cmd='rcp -t /tmp/'
Feb 18 12:27:10 node1 in.rshd[4094]: connect from 212.202.154.173 (212.202.154.173)
Feb 18 12:27:10 node1 pam_rhosts_auth[4094]: allowed to root@node1.moinmoin.de as root
Feb 18 12:27:10 node1 in.rshd[4095]: root@node1.moinmoin.de as root: cmd='rm /tmp/node1.scnlg'
Feb 18 12:27:10 node1 in.rshd[4102]: connect from 212.202.154.173 (212.202.154.173)
Feb 18 12:27:10 node1 pam_rhosts_auth[4102]: allowed to root@node1.moinmoin.de as root
Feb 18 12:27:10 node1 in.rshd[4103]: root@node1.moinmoin.de as root: cmd='rm /tmp/node1.cfg'


But, when I try to create the Cluster configuration file I am getting the following error,

------------------------------------------------------------

node1:/home/admin # cmquerycl -v -L /dev/cciss/c0d1p1 -n node1 -n node2 -C $SGCONF/lpcluster.config

Begin checking the nodes...
Error: Permission denied to 212.202.154.172
Looking for other clusters ... Done
Error: Unable to establish communication to node node2: 19
Failed to gather configuration information.
node1:/home/admin #
------------------------------------------------------------

The log output of the "node1" is,
------------------------------------------------------------

Feb 18 12:30:18 node1 cmclconfd[4144]: Unable to stat /opt/cmcluster/conf/cmclconfig, No such file or directory
------------------------------------------------------------

there is no cmclconfig file under /opt/cmcluster/conf/ directory.

What could be the problem?

[Serviceguard for Linux]

First thing to try is to add the IP addresses to the cmclnodelist file.

From the manual (managing pg 124)


In the following example, NodeA and NodeB share two public subnetworks, 192.6.1.0 and 192.6.5.0. NodeA is configured in subnet
192.6.1.0 and its official IP address known by clients is 192.6.1.10. It also
has another IP address, 192.6.1.20, which is available for communication on subnet 192.6.5.0.

The cmclnodelist file must contain all these IP addresses in order to grant permission for ServiceGuard messages to be
sent from any node.

NodeA root # cluster1
192.6.1.10 root # cluster1
192.6.5.20 root # cluster1
NodeB.sys.dom.com root # cluster1
192.6.1.12 root # cluster1
192.6.5.22 root # cluster1

If that does not work, try to bring up a single node cluster and report on the results.