LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: Failover with NAT RH 6.2 & UM 1.0.1 (Kernel 2.2.14-5.0.14.um.3)

To: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Subject: Re: Failover with NAT RH 6.2 & UM 1.0.1 (Kernel 2.2.14-5.0.14.um.3)
Cc: Shaun McCullagh <smccullagh@xxxxxxxxxxxxxxxxxxxxxxxxxx>
From: Horms <horms@xxxxxxxxxxxx>
Date: Tue, 17 Oct 2000 12:14:20 -0400
On Tue, Oct 17, 2000 at 05:59:30PM +0200, Juri Haberland wrote:
> Shaun McCullagh wrote:
> > 
> > Juri,
> > 
> > Thank you very much indeed!
> > 
> > I had to put
> > 
> > fizz.information-innovation.com IPaddr::192.168.31.157
> > ldirectord::ldirectord.cf
> > 
> > into haresources, rather than
> > 
> > fizz.information-innovation.com 192.168.31.157 ldirectord
> > 
> > (ldirectord exited with code 1)
> 
> Hm, I thought ldirectord would search for ldirectord.cf somewhere in
> /etc/ha.d/conf if it is not told otherwise - but never mind.
> What I don't understand is why you are telling heartbeat explicitly to
> use the IPaddr script. It would do so also if you just put your ip
> address in the haresources file - but there's nothing wrong with what
> you are doing. It's just that I wouldn't :-)

ldirectord has to be told which configuration file to use 
(e.g.  ldirectord.cf). It will look for this configuration file
in /etc/ha.d/conf

> > but now it works purrfect :)
> 
> Great :-)
> 
> > One last question:
> > 
> > If each LVS machine has one ppp interface, don't they need unique IP
> > numbers?
> 
> Ehem. Right.
> 
> > You mentioned that haresources, ha.cf and authkeys should be identical
> > on both machines.
> 
> Well, actually I never used the ppp-udp stuff. I always used plain
> serial communication. It worked for me and the ppp stuff looked too
> troublesome to me. In my case you can just copy the files to the other
> node, but in your case you indeed have to maintain different versions of
> ha.cf.

Ditto.

I have always used ethernet-udp and/or raw serial but not PPP.

-- 
Horms


<Prev in Thread] Current Thread [Next in Thread>