So it finally succeeded in loading all the mailboxes with a beefier server.
The problem was, I think, that with 60.000 mailboxes the first server ran out of memory and began swapping. A swapping server seldomly does any good.
Next problem I ran into is the "An entry with the same key already exists" which I see there is a thread on already and fix coming.
It is still slow, though. Loading all the boxes takes 15-30 MINUTES. This on a server with 32 GB RAM and a 10 GbE internet connection.
There really ought to be some sort of filtering option when selecting mailboxes or creating jobs dynamically based on filter criterias.
This could be accomplished with one of the following:
- Mailbox membership (possibly problematic?; mailbox can be member of multiple groups)
- Retention policy
- Custom attributes
Look, it became a feature request!