Description of problem: When using 2-node clusters and /etc/hosts is improperly configured in such a way that the local node name resolves to a loopback address, a split-brain can ensue. Thus, it would be nice if, when starting, CMAN could check for the improper use of loopback and abort/not run if this occurs (as well as printing a loud warning). While the case where this would cause problems in production is probably astronomically low (this is not a BUG, it's a feature request), it does and has caused confusing for new installations when configuring the cluster for the first time.
...it does and has caused confusion when configuring the cluster for the first time. Making sense is a good thing.
Yes... my installation of RHEL 4 came up with my hostname in /etc/hosts at the loopback address that came up with the cluster confusion. samples on node1 config sample: notworking 127.0.0.1 mynode1 localhost.localdomain localhost 192.168.0.105 mynode2 192.168.0.104 mynode1 #added by myself since i had probs with the cluster config sample fixed & working 127.0.0.1 localhost.localdomain localhost 192.168.0.105 mynode2 192.168.0.104 mynode1 good thing would me a note in syslog or something like that greeeeetz markus.jeckeln
Added to CVS head and FC4 branches.
Added to RHEL4 branch so it should appear in U2