Full-Text Search
Overview
Full-text search allows all messages to be indexed, so that mail clients can quickly and efficiently search messages by their full text content. Dovecot supports a variety of community supported FTS indexing backends.
docker-mailserver
comes pre-installed with two plugins that can be enabled with a dovecot config file.
Please be aware that indexing consumes memory and takes up additional disk space.
Xapian
The dovecot-fts-xapian plugin makes use of Xapian. Xapian enables embedding an FTS engine without the need for additional backends.
The indexes will be stored as a subfolder named xapian-indexes
inside your local mail-data
folder (/var/mail
internally). With the default settings, 10GB of email data may generate around 4GB of indexed data.
While indexing is memory intensive, you can configure the plugin to limit the amount of memory consumed by the index workers. With Xapian being small and fast, this plugin is a good choice for low memory environments (2GB) as compared to Solr.
Setup
-
To configure
fts-xapian
as a dovecot plugin, create a file atdocker-data/dms/config/dovecot/fts-xapian-plugin.conf
and place the following in it:mail_plugins = $mail_plugins fts fts_xapian plugin { fts = xapian fts_xapian = partial=3 full=20 verbose=0 fts_autoindex = yes fts_enforced = yes # disable indexing of folders # fts_autoindex_exclude = \Trash # Index attachements # fts_decoder = decode2text } service indexer-worker { # limit size of indexer-worker RAM usage, ex: 512MB, 1GB, 2GB vsz_limit = 1GB } # service decode2text { # executable = script /usr/libexec/dovecot/decode2text.sh # user = dovecot # unix_listener decode2text { # mode = 0666 # } # }
adjust the settings to tune for your desired memory limits, exclude folders and enable searching text inside of attachments
-
Update
docker-compose.yml
to load the previously created dovecot plugin config file:version: '3.8' services: mailserver: image: docker.io/mailserver/docker-mailserver:latest container_name: mailserver hostname: mail domainname: example.com env_file: mailserver.env ports: - "25:25" # SMTP (explicit TLS => STARTTLS) - "143:143" # IMAP4 (explicit TLS => STARTTLS) - "465:465" # ESMTP (implicit TLS) - "587:587" # ESMTP (explicit TLS => STARTTLS) - "993:993" # IMAP4 (implicit TLS) volumes: - ./docker-data/dms/mail-data/:/var/mail/ - ./docker-data/dms/mail-state/:/var/mail-state/ - ./docker-data/dms/mail-logs/:/var/log/mail/ - ./docker-data/dms/config/:/tmp/docker-mailserver/ - ./docker-data/dms/config/dovecot/fts-xapian-plugin.conf:/etc/dovecot/conf.d/10-plugin.conf:ro - /etc/localtime:/etc/localtime:ro restart: always stop_grace_period: 1m cap_add: - NET_ADMIN
-
Recreate containers:
docker-compose down docker-compose up -d
-
Initialize indexing on all users for all mail:
docker-compose exec mailserver doveadm index -A -q \*
-
Run the following command in a daily cron job:
Or like the Spamassassin example shows, you can instead usedocker-compose exec mailserver doveadm fts optimize -A
cron
from withindocker-mailserver
to avoid potential errors if the mail-server is not running:
Example
Create a system cron file:
# in the docker-compose.yml root directory
mkdir -p ./docker-data/dms/cron # if you didn't have this folder before
touch ./docker-data/dms/cron/fts_xapian
chown root:root ./docker-data/dms/cron/fts_xapian
chmod 0644 ./docker-data/dms/cron/fts_xapian
Edit the system cron file nano ./docker-data/dms/cron/fts_xapian
, and set an appropriate configuration:
# Adding `MAILTO=""` prevents cron emailing notifications of the task outcome each run
MAILTO=""
#
# m h dom mon dow user command
#
# Everyday 4:00AM, optimize index files
0 4 * * * root doveadm fts optimize -A
Then with docker-compose.yml
:
services:
mailserver:
image: docker.io/mailserver/docker-mailserver:latest
volumes:
- ./docker-data/dms/cron/fts_xapian:/etc/cron.d/fts_xapian
Solr
The dovecot-solr Plugin is used in conjunction with Apache Solr running in a separate container. This is quite straightforward to setup using the following instructions.
Solr is a mature and fast indexing backend that runs on the JVM. The indexes are relatively compact compared to the size of your total email.
However, Solr also requires a fair bit of RAM. While Solr is highly tuneable, it may require a bit of testing to get it right.
Setup
-
docker-compose.yml
:solr: image: lmmdock/dovecot-solr:latest volumes: - ./docker-data/dms/config/dovecot/solr-dovecot:/opt/solr/server/solr/dovecot restart: always mailserver: depends_on: - solr image: docker.io/mailserver/docker-mailserver:latest ... volumes: ... - ./docker-data/dms/config/dovecot/10-plugin.conf:/etc/dovecot/conf.d/10-plugin.conf:ro ...
-
./docker-data/dms/config/dovecot/10-plugin.conf
:mail_plugins = $mail_plugins fts fts_solr plugin { fts = solr fts_autoindex = yes fts_solr = url=http://solr:8983/solr/dovecot/ }
-
Recreate containers:
docker-compose down ; docker-compose up -d
-
Flag all user mailbox FTS indexes as invalid, so they are rescanned on demand when they are next searched:
docker-compose exec mailserver doveadm fts rescan -A
Further Discussion
See #905