kolab v1 -> kolab v2

[Gavin McCullagh]

Hi Thomas,

thanks for your help...

On Fri, 05 Feb 2010, Thomas Arendsen Hein wrote:

> > Our current 1st.README only details the required instruction to upgrade 
> > from 2.1.0 onwards. I'd suggest to search the mailing list for further 
> > instructions to upgrade from 1.0.
> 
> Manual import of all data might be the easiest.

I've been working this and have a script which

 * runs slapcat on kolabv1
 * a perl script goes through and makes the necessary modifications,
   including splitting out users into multiple domains.
 * dumps the mailbox list on kolab v1, translates it and loads it up in the
   kolab v2
 * rsyncs the users mailbox into the right new username and domain-specific
   location 
 * reconstructs the user's mailboxes
 * transfers over user.seen and users.sub (though the latter is probably
   not necessary?

I'm not quite there yet, but it seems to work reasonably so far.

> >> The location and structure of spools has changed a bit with the inclusion
> >> of domains though it's not too tough to spot where things are -- although I
> >> also don't see the sieve directory as yet.
> 
> You can configure the layout, searchfor hashimapspool in 1st.README

I think I'd prefer to just massage our data into the standard layout.  I
reckon if we're as consistent as possible with default Kolab v2.2, things
will be easier in future.

Gavin