On Thu, 4 May 2006 18:55:13 +0100 Peter Curran <[EMAIL PROTECTED]> wrote:
> I can confirm Scott's words - I was one of the people with the problem. > > My results where exactly the same as you are seeing (except mine was only on > one interface). Carp multicasts stuff on the interface at regualr intervals > and expects to see multicasts from the other system coming in. It is the > frequency of these multicasts that determines who is master, and who slave. > > The advskew 100 simply slows down the multicast rate of the slave. > > The fact that both systems are master means that they are not seeing each > others multicasts. > > In my case the problem was caused by having a switch that was divided into > different VLANs (by port). The two systems where plugged into the same > switch, but not into the same VLAN. Once I had discovered this (by plugging > them into the correct ports) the problem was fixed instanltly. > > Cheers > > Peter > In my case there is no switch between the two CARP boxes, but a Xover cable. But I can suspect the cards, which are old 3Com ISA bus cards (ep) because there are no more free pci bus on the board. I will try to swap them tomorrow with one of the pci's. Thanks Imre -- This is Linux Land. In silent nights you can hear the windows machines rebooting
signature.asc
Description: PGP signature