I found a tool in the procmail suite to split mbox archives into
individual messages. I upped the available swap space on the server to
13GB (thanks to David Kstrup for suggesting this idea). There are 156457
items in the archive up to last month or so. That number is from the
first message ever posted to now. The import went fine. Only took a few
hours.
The import creates what are called staged users, not yet activated. I
need to learn how to let people activate their accounts or automate that.
This appears to be a complete success. The only surprise is finding some
emails right at the top that are junk about cheap DVD burners, complete
spam, and without proper headers. Had me perplexed for a while, but you
can actually see this junk if you search in the user archives.
When I play some more with this instance I'll open it up for everybody
to have a look at if they would like. I'll publish the domain name that
it sits on then.
One thing I like about Discourse is the good search function. Sure you
can search the web based user archive but this is all integrated on the
Discourse web interface and it's a pleasure to use. The recent versions
of Discourse also have Chat, which is a potentially useful feature.
Andrew
- Discourse experiments Andrew Bernard
-
