I'm running SpamAssassin 2.63 with MailScanner 4.26.8-1 under Tru64 UNIX 5.1B. MailScanner has all of our Bayes files stored in /var/spool/spamassassin. Having one systemwide Bayes database is new to me but I'm noticing TONS of files named bayes_toks.expirennnnn where nnnnn is some number. What is the purpose of all of these files? They're consuming massive amounts of disk space. Thanks in advance.
If I'm not mistaken, mailscanner 4.26 doesn't do bayes expiry management like 4.28 does.
What I'm guessing is going on is whenever mail is handed to SA, it's trying to do an opportunistic expiry, because you are past-due for one. Mailscanner gives up on SA, times it out and kills it before the expiry can complete, leaving the temporary file behind.
what to do to fix things in the short-term:
1) manually run sa-learn --force-expire right now
2) clean up the old expirennnn files.
3) run sa-learn --force-expire as a daily cron job. Be sure this is done in the crontab for your MailScanner user, and not some arbitrary user.
In the long term, I'd suggest using mailscanner 4.28.. it manages bayes expiry directly to prevent this kind of trouble.
