==== REQUIRED BASIC INFO OF YOUR IREDMAIL SERVER ====
- iRedMail version (check /etc/iredmail-release): 1.6.8
- Deployed with iRedMail Easy or the downloadable installer? Installer
- Linux/BSD distribution name and version: Rocky 9.3
- Store mail accounts in which backend (LDAP/MySQL/PGSQL): MariaDB
- Web server (Apache or Nginx): Deprecated issue!
- Manage mail accounts with iRedAdmin-Pro? No
====
Just posting to be sure I understand my options about spam learning properly. I understand there to be three aspects:

1. Set up SpamAssassin to have an account for bayes data in an SQL database as per https://docs.iredmail.org/store.spamass … n.sql.html

2. Set up auto-learning (setup imap_sieve to duplicate messages specifically accepted / rejected into special folders and use cron to regularly run the bayes classifier) as per https://docs.iredmail.org/dovecot.imapsieve.html

3. Find a representative collection of ham and spam, then manually run sa-learn to (better) teach SpamAssassin the difference.

My questions are:
- is 1 a prerequisite for 2 and / or 3? Can I use only 3 (i.e. sa-learn with no SQL account)?
- with 2, is 1 mandatory, optional, or prohibited?
- can I use all 3 (auto-learn generally, augmented by additional manual training)?
- is something missing from the above list?

Thanks for all help as always,
Paul