[plug] Dovecot FTS

[Byron Hammond]

Cool. That sounds really good Brad. I've heard of Tika before and have used
it through my use of paperless ngx, which you may want to take a look at
for document storage, searching etc

On Tue, 25 Aug 2026, 9:58 pm Brad Campbell via plug, <plug at plug.org.au>
wrote:

> G'day All,
>
> I've run Zimbra for the office since about 2011. One of the things I love
> about Zimbra is the full text search. It indexes everything no matter what
> it's buried in (office docs, libreoffice docs, pdf, zip, whatever).
>
> I run Exim/Dovecot for my personal stuff with Thunderbird as a client.
> Last year I had a bit of a look around to see what might work with Dovecot
> and was pleased to see there are a number of supported solutions. I had a
> play on my staging system and had settled on a combination of Apache Solr
> and Tika. Then when I went to try it out I noted Dovecot had in general
> moved away from Solr to flatcurve. Bonus. Solr had quite an infrastructure
> around it while flatcurve just puts its indexes in the maildirs, so no sync
> issues and "just works" for backups.
>
> I set up a VM to poke Tika in because I had already tried out Solr in the
> VM and it was there. About 10 minutes of config and an hour or so to index
> 20 years of mail and everything was done.
>
> So having settled on a config I upgraded the main server today and rolled
> it out. Set Thunderbird up to disable its indexer (which sucks for large
> mail spools anyway) and search on-server.
>
> I'm still having issues with the server (periodic MCEs), so a good test
> was to dig out the pdf receipts for the RAM and CPU. Popped
> "CMK64GX4M2E3200C16" and 3 seconds later up came the pdf buried in a folder
> I'd "Safely filed for later".
>
> Being able to do body and attachment searches on >50K e-mails in seconds
> is awesome. I used to run a search then get a coffee while Thunderbird did
> a body text search, only to not get what I was looking for because it was
> buried in a pdf. No longer!
>
> Tika is super simple to set up (literally download the jar and run it from
> java). It's also a lot lighter weight than I was prepared for. Now to roll
> it out at mums place for her mail archive.
>
> Regards,
> Brad
> _______________________________________________
> PLUG discussion list: plug at plug.org.au
> https://lists.plug.org.au/mailman/listinfo/plug
> Committee e-mail: committee at plug.org.au
> PLUG Membership: http://www.plug.org.au/membership
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.plug.org.au/pipermail/plug/attachments/20260826/4dc3a4ed/attachment.htm>