Do you have 'preemption' enabled on both boxes in the cluster? This is the trigger for CARP to take down all CARP interfaces if any one goes down (for whatever reason - I see this at work on OpenBSD from time to time).
--Bill On 10/23/05, Britton Kerr <[EMAIL PROTECTED]> wrote: > Since .88 I'll have one of my interfaces go backup while the others stay > master. There are no cable connection problems. Once I disable carp on the > master and reenable and sometimes I have to do the same on the backup the > master will reassume on all interfaces. Problems arise when users have apps > open and this happens. All carp interfaces should failover if one interface > goes down (or is perceived as down). There are actually two issues here. > 1. Carp fails over on an interface when I can still ping it from the > network...ie it was never really down and 2. if one fails they all should at > the same time to keep the connections uninterrupted. > This has been occuring about 3-4 times last week with with .88 (I disabled > carp on Wed to keep it from happening anymore and it happened overnight > after upgrading to 89.2. What will be the best way help diagnose the > intermittent failover problem? Grab log file + screen shots before or after > I get it working again? And which files? > I'm running on production network with 30 or so machines now and will have > 250 machines eventually behind them. > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
