RE: Internal virtual IP failure scenario

Thanks, though I've run into a problem that seems to be specific to the
ping_group directive.

The ping_group I set up with the real servers always shows dead. However
I can ping the Real Servers from each Director. Adding them individually
as just ping targets instead of a group in ha.cf works fine.

I found a thread that claims ping_group is broken in heartbeat 1.2.4. Is
it broken in the 1.2.3 CVS version UM3 appears to be using as well?

-----Original Message-----
From: Horms [mailto:[email protected]] 
Sent: Monday, July 03, 2006 6:01 PM
To: Bryn Moslow
Cc: [email protected]
Subject: Re: Internal virtual IP failure scenario

On Mon, Jul 03, 2006 at 12:29:38PM -0700, Bryn Moslow wrote:
> Thanks, yes I am pinging the router on the public IP side for ipfail
in
> my setup. I was hoping to not need to use a Real Server as a ping
test.
> If the Real Server dies, I believed this would simply result in
> ldirectord taking the unreachable Real Server out of the equation but
no
> fail over. Is this correct, or will the setup be smart enough to see
> that the inactive Director is able to reach the Real Server in
question
> and trigger a fail over?
> 
> It just struck me squarely between the eyeballs that I guess I can
fire
> the cluster back up and test this on my own. "Duh" for lack of a
better
> expression...
> 
> Maybe this would be a good/better use of ping_group as well, just
adding
> the list of Real Servers there.

If you are are unable to ping something a bit more solid on the
real-server side, like say the switch, then a ping_group that includes
all the real servers would be a good second choice.

-- 
Horms                                           
  H: http://www.vergenet.net/~horms/
  W: http://www.valinux.co.jp/en/



-- 
Ultra Monkey - http://www.ultramonkey.org/
To UNSUBSCRIBE, email to [email protected], with a body:
unsubscribe ultramonkey-users [email protected]
where "[email protected]" is YOUR email address.