Hi David, On Wed, Dec 08, 2010 at 06:30:21PM -0500, David Birdsong wrote: > On Wed, Dec 8, 2010 at 6:13 PM, Bryan Talbot <btal...@aeriagames.com> wrote: > > See the architecture doc section 4.3 > > http://haproxy.1wt.eu/download/1.3/doc/architecture.txt > > -Bryan > > > When a new haproxy pid is started after it sends a SIGTTOU to the > prior running haproxy pid, what is the state of the backends? If one > backend was in a down state due to failed health checks, will it be > marked up by the new process until it fails the configured number of > checks?
They're marked UP for one single check (as if they failed all checks but the last one). That way, the first check fixes the status. Regards, Willy
