I have 3 servers running centos 5.2 and running kernel 2.6.18-92.1.6.el5 x86_64. One of the servers i am only using for monitoring and managing the cluster (with conga/luci). I have a private lan setup on eth1 just for managing the cluster and for connecting to the iscsi SAN. The apf firewall is setup to trust all traffic on eth1.

Anyway, i get the following error after creating the cluster with the two nodes. This error shows in the /var/log/messages file on each server.

Jul 3 03:35:51 skydive openais[5628]: [MAIN ] Using default multicast address of 239.192.113.184
Jul 3 03:35:51 skydive openais[5628]: [MAIN ] Error reading config from CCS
Jul 3 03:35:51 skydive openais[5628]: [MAIN ] AIS Executive exiting (reason: could not read the main configuration file).
Jul 3 03:36:16 skydive ccsd[5595]: Unable to connect to cluster infrastructure after 30 seconds.


Ive tried googling for answers or people having the same type of issue, but I have came up empty. Is it trying to read cluster.conf? If so, is it the one on the luci server or on itself? Note: the skydive server is one of the cluster nodes, not the server that is running Luci.

Any ideas?

Looks like everyone else is either to busy or just as long as i am. I am hoping to get some ideas to resolve this without having to hire a consultant. =/

check your /etc/hosts file, make sure that 127.0.0.1 does not use the same name as your node.

If you try to start cman, this is probably the error you will get.

cman not started: Local host name resolves to 127.0.0.1; fix /etc/hosts before starting cluster. /usr/sbin/cman_tool: aisexec daemon didn't start

just change it to localhost

eg.

127.0.0.1 localhost.localdomain localhost