Hi Rob,
My comment was meant as supporting evidence that this task is doable and not blocked by possible security issues.
But I now have experience with using Confluence to bust apart 10 years of support email and am going to list a few other issues that you will encounter if you are also trying to do the same thing.
First, it takes forever to import email into Confluence. On a modern 2.66GHz quad proc Xeon with 24GB of RAM, it took nearly an hour to import last year's support archive of 4500 emails. Conversely, it took mhonarc less than two minutes to process the same archive.
Once the email is inside Confluence, the search functionality works quite nicely. However, everything, both the email messages and the attachments, are displayed in search results with the date you fed the archive in. Once you open the email, the right date is displayed, but when the email shows up in a search list, the feeding day appears.
I spent the better part of a day feeding in our entire 10 year support email list. But this issue of only showing mail in plaintext was just too painful, so I decided to try and delete the Space that contained the archive. I thought this would go very quickly, but I think the system tried to delete one email at a time. There were 25,000 of them, so this was taking forever. I waited a couple of hours and didn't see any further progress (disk space was no longer decreasing, but CPU was pegged), so I just gave up and deleted my entire Confluence installation and re-installed it.
I would not recommend Confluence for large email archives even if they fix this HTML display issue because of these other problems I have encountered.
mhonarc isn't perfect and it is missing a search engine that you have to add separately (however, the Namazu engine seems to be built for it). But it has the advantage of being both highly configurable without changing the code, and open source. And it is blazing fast - you can build your 10 year mail archive in under 15 minutes.
Good luck no matter which direction you go.
todd
Hi Everyone, after being open for so many years and with little traction, I'm closing this issue as we cannot see ourselves implementing this in the foreseeable future.