On Mon, Apr 23, 2007 at 10:18:27PM -0500, isplist@xxxxxxxxxxxx wrote:
> >> ]# ipvsadm -L -n
> >> IP Virtual Server version 1.2.0 (size=4096)
> >> Prot LocalAddress:Port Scheduler Flags
> >> -> RemoteAddress:Port Forward Weight ActiveConn InActConn
> >> TCP 192.168.1.151:143 wlc
> >> TCP 192.168.1.151:25 wlc
> >> TCP 192.168.1.151:80 wlc
> >> -> 192.168.1.250:80 Route 1 0 0
> >> -> 192.168.1.249:80 Route 1 0 0
> >> TCP 192.168.1.151:110 wlc
> >> Actually, that is very interesting. It shows exactly what's happening.
> >> Only port 80 traffic has been able to get through, none of the others.
> > In that case, is http working ?
> Yes, so are all of the other services. If I connect directly to the mail
> servers, all services are running. If I connect through LVS, only http makes
> it, everything else fails.
Http working is a good sign.
> > There are no real servers assigned to the other virtual services.
> > You are using keepalived, right? I guess its health checker thinks
> > that the real servers for the non-http virtual services are unavailable.
> > Perhaps because they aren't responding as keepalived expects.
> Keepalived, no, I've always thought that was part of another project. I'm only
> using LVS/pulse. The other LVS servers work fine for web services.
Ok, I was just guessing. In that case, the pulse health checks don't
seem to think that your real servers are up.