Hey Hey,

I'm just looking through the Linux-ha to provide a reliable floating IP
between a master web-server and slave fail-over. unfortunately there are a
couple of issues. This set-up is replicated about 600 times all on a very
out of date source based distro where upstream support has disappeared.
Just to give you an Idea, I had to upgrade python to meet the minimum
requirements for pacemaker!! While we are in the process of moving this to
a more up to date system, We're gonna be stuck with the Old machines for a
while.

Is there a way around the node directive values needing to match uname -n,
could I use for example an IP, or system uuid etc?

The ha.cf requires either an explicit node configuration or a autojoin any.
As this is going to be running on a network I have no control of, and will
only have 2 and only 2 nodes, I am reluctant to have the autojoin any
option.

So nodes it is, which is where the issue arises. from the ha.cf man page
"Node names in the directive must match the "uname -n" of that machine".
Part of the horrible system I have inherited means that all boxes have the
same host name. Which causes some issues when they both try to join the
cluster. While changing the hostname shouldn't be a massive deal, in this
case it is more than a simple issue as there are a number of configurations
that rely on it being what it is. I can go in and have a hunt for bits of
code that require the specific hostname and try to rectify the issue, but
it's a lot easier to let sleeping lions lie whilst we get the product
running on a Centos host and start shipping that instead!

Any pointers or help would be much appreciated.

Many Thanks

Pete Uttley
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to