On 11/5/07, Sloan <user-b1d2c84d244b@xymon.invalid> wrote:
Haha, what I mean is, big brother is showing it's age, and getting
harder to support as time goes by, simply by virtue of the fact that the
code is slowly rotting. But we can't just make a clean sweep - so much
now depends on the way big brother behaves, that the replacement needs
to be able to act identically to big brother.
In your original email you said:
"only one of the bb servers does reporting, as determined by the
failover state"
Does that mean only one of the two sends out notifications?? If so,
why not just have two identical Hobbit servers, with all the clients
sending reports to both, and disable the [bbpage] section in
server/etc/hobbitlaunch.cfg on the backup system. Then your failover
solution only has to detect "death of twin" and rewrite the
hobbitlaunch.cfg with DISABLED removed from that section. I guess
Hobbit might need to be reloaded or restarted for the change to take
effect.
I haven't tried this - we don't use the paging system - I'm just
wondering if something as simple as that might work.
Ralph Mitchell