|
On Tue, 15 Jul 2003, Horms wrote:
> On Tue, Jul 15, 2003 at 10:05:49AM +0200, Stephan Wonczak wrote:
> >
> > If some client machine has to do a new arp-request, sometimes the now
> > secondary machines answeres it! Meaning: Machine B is director, having
> > taken over service from machine A, but both are still running. This
> > happens e.g. during maintenance. Machine B has both :0 and :1-Adresses,
> > machine A does no longer (verifiable by ifconfig).
> > Using tcpdump we could see machine A still answering arp-requests for
> > the public LVS-Address, even though it is now assigned to machine B who
> > *should* be answering. Huh? The funny thing is, this migration of IP
> > adresses on virtual interfaces forks just fine without this problem for
> > numerous other services, only ipvs seems to produce the problem with the
> > ARPs.
>
> Is the IP address still attached to an interface (alias or otherwise)
> in Linux Director A?
No, at least not in any way we can determine. The aliased interfaces are
brought down correctly by the cluster software; this works for all other
non-LVS-services without ony problem. Only ipvs seems to do something
different; at least *something* is left behind that causes the unwanted
ARP answers. It would be very helpful to have a pointer on what to check
since we are fresh out of ideas :-)
Thanks for your help!
Dipl. Chem. Dr. Stephan Wonczak
Institut fuer Angewandte Informatik (ZAIK)
Regionales Rechenzentrum der Universitaet zu Koeln (RRZK)
Universitaet zu Koeln, Robert-Koch-Strasse 10, 50931 Koeln
Tel: ++49/(0)221/478-5577, Fax: ++49/(0)221/478-5590
|
