El jue, 22-01-2004 a las 17:58, Chris Shenton escribi�:
> > Of course we would also have to deal with errors on the connection with
> > the backup, and provide some kind of mechanism (another queue?) to
> > re-sync the backup whenever there's been a network glitch.
> Yeah, that gets difficult. Plenty of challenges here.  :-)

The clustering code should already have a mechanism (queue) to deal with
SMTP (e.g. a mail for server B gets to server A, but B is down). We
would need another queue to propagate the deletions to the backup. 

In the best case, this code should also be able to "clean" the queue,
something like this:

- mail gets to the primary server
- the backup is down so it gets queued
- a user reads and deletes this mail
- the server detects that the backup is down and this particular mail is
still in the backup-syncing queue, so it deletes it from the queue

Shouldn't be that difficult as every mail has a unique id: the Maildir
filenames of every message are unique, aren't they?

-- 
 Vicente Aguilar <[EMAIL PROTECTED]>
 Departamento de Sistemas
 Tlf.: 965 98 71 92

 Recursos en la Red, S.L.U.
 http://www.renr.es

Reply via email to