Patric Falinder:
> > as a result, so I guess it doesn't have to do with the timer thing!?
No you cannot prove the absence of this bug. Perhaps it triggers
when you run lots of instances of my test program in parallel. We
have no idea of what all the other virtual machines are doing and
we have no idea how VMware tools futzes with the guest system clock.
> > I also upgraded to Postfix 2.7, without VDA cause this time because this
> > server wasn't meant to have that patch in the first place, and I've been
> > watching the queue for a couple of minutes and the double-bounce
> > messages seems to go away as they should.
> > I'm gonna keep looking though, to see if it gets filled again but it
> > Looks like the upgrade fixed it.
> > I can report back at the end of the day to see if it's actually fixed.
As I wrote before there are no changes in this part of Postfix
since version 2.4.
I expect that lost timer bugs will come back again and again. This
weekend someone else reported a lost timer interrupt problem in
FreeBSD on ESX that caused idle Postfix daemons to hang instead
of exiting after $max_idle seconds.
Wietse
