replication is unable to fix uidvalidity difference

[Bron Gondwana]

On Mon, 3 Sep 2007 10:24:43 +0100 (BST), "David Carter" <dpc22 at cam.ac.uk> said:
> On Mon, 3 Sep 2007, Bron Gondwana wrote:
> 
> > My reading was a reconstruct without cyrus.index.  If you reconstruct 
> > without a header it will get a new UniqueID, but if you reconstruct 
> > without an index it doesn't know the UidValidity, so it sets it time().
> 
> Agreed. Unfortunately, that's a rather more common failure mode.
> 
> I'll have a think about this after I work out why shutting down Cyrus 
> causes my skiplist mailboxes.db to explode one time in three. Didn't used 
> to happen with 2.1.X, and I haven't touched the skiplist stuff :(.

happened for the first time ever to me this afternoon.  Most annoying.

Thankfully the replica was entirely up-to-date and I could just copy the
one from there, since our automatic restart script managed to restart it
and overwrite the two backup copies we keep across 2 restarts.  Oops.

We get a few errors with skiplists still, but hopefully clean shutdown
will help with that.

Bron.
-- 
  Bron Gondwana
  brong at fastmail.fm