Calling scp /etc/my.cnf node2:/etc/ on a node which is part of a Corosync cluster gets annoying after a while. That command has also redundant information in it:
[root@node1 ~]# clucp /etc/hostsAdd -v to see what's happening:
[root@node1 ~]# clucp -v /etc/hosts sending /etc/hosts (158 B) to 1 node node 673753280 (192.168.40.168): OKYou can copy multiple files at a time:
[root@node1 ~]# clucp -v /etc/hosts /etc/resolv.conf sending /etc/hosts (158 B) to 1 node node 673753280 (192.168.40.168): OK sending /etc/resolv.conf (96 B) to 1 node node 673753280 (192.168.40.168): OKList all nodes which have joined the cluster and started clucpd:
[root@node1 ~]# clucp -l I group node id pid addrs * clucp 36219072 1601 192.168.40.167 clucp 673753280 2116 192.168.40.168Hint: A star in the column labeled I denotes the node you're currently working on.
Take either an RPM from below or compile clucp on your own. You'll need a C compiler, make, and corosynclib-devel for that:
tar -xjf clucp-0.1.tar.bz2 cd clucp-0.1 make sudo make installThe RPM installs binaries below /usr, whereas the source package installs them below /usr/local unless you call make install with a prefix:
sudo make prefix=/opt/clucp install