Expired Zone "Recovery" Process

[Barry Margolin]

In article <e6qtmc$2b8r$1 at sf1.isc.org>, colin.kopp at gmail.com wrote:

> Barry Margolin wrote:
> > In article <e6pi4n$2tev$1 at sf1.isc.org>,
> >  "Shaheen" <wael.shaheen at gmail.com> wrote:
> >
> > > Actually am not sure if this can be done in other ways or it is only
> > > the behavior of BIND,
> > > when the master goes down and the slave expires the zone you will not
> > > get answers from either DNS servers,
> > > to fix this i used to delete all records and recreate them again from
> > > both master and slave servers as soon as i get the master DNS up again.
> >
> > I've never had that experience.  When the master came back, the slave
> > transferred the zone from it and started responding normally.
> >
> 
> So this would seem to imply that when the master is restarted it will
> send out a NOTIFY to the slaves and even though the slaves have
> expired the zone they will action the NOTIFY.  Is that a fair
> statement?

The master doesn't know that the slaves have expired the zones.  Masters 
send out notifies to all slaves every time they start up.

> What happens in the case where there was a network connectivity issue
> between the master and the slave.  In this instance the master has no
> need to send a NOTIFY because it has been up and running and has
> not been aware of the loss of connectivity.  I presume that
> a manual reload of the zone on the slave would be necessary.
> Am I on the right track here?

In this case the retry timer on the slave causes it to poll the master 
periodically.  I don't think this stops when the slave expires the zone.

-- 
Barry Margolin, barmar at alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***
*** PLEASE don't copy me on replies, I'll read them in the group ***