You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Bill Landry <bi...@pointshare.com> on 2004/10/23 19:56:35 UTC
Bayes sa-learn change in SA 3.0.1?
I have "bayes_auto_expire 0" in my local.cf and instead run
"sa-learn --force-expire" via an hourly cron job. With SA 3.0.0 I was see
the following results in my hourly cron results e-mails:
=====
synced Bayes databases from journal in 3 seconds: 3677 unique entries (3714
total entries)
expired old Bayes database entries in 83 seconds
129678 entries kept, 10414 deleted
token frequency: 1-occurence tokens: 55.30%
token frequency: less than 8 occurrences: 26.17%
=====
Since upgrading to SA 3.0.1 last night, I am now only seeing the following
in my hourly cron results e-mails:
=====
synced Bayes databases from journal in 3 seconds: 3268 unique entries (3478
total entries)
=====
I manually ran "sa-learn --force-expire -D" to see what was going on, and am
seeing the following lines in my debug output:
=====
debug: bayes: Can't use estimation method for expiry, something fishy,
calculating optimal atime delta (first pass)
debug: bayes: couldn't find a good delta atime, need more token difference,
skipping expire.
=====
Is this new behavior to be expected now with SA 3.0.1?
Bill
Re: Dnsbl lookups not working since upgrade to 3.0.0
Posted by Niek <ni...@packetstorm.nu>.
On 10/23/2004 11:28 PM +0200, marti wrote:
> I am running suse 8.1 just upgraded from 2.64 everything works bar the dnsbl
> lookups, not had one positive result since upgrading.
> Is there some other perl modules I need to upgrade? Other than the
> perl-spamassassin-3.0.0-1.i586.rpm
>
> Martin
Please don't hijack threads.
Meaning: Don't reply to an old message from the list, and change the
subject. This screws things up.
That being said,
Is your Net::DNS up to date?
If so, and still doesn't work have a look here:
http://www.surbl.org/faq.html#nettest
Regards,
Niek
--
_______________________________________________________________________
Read about mime: http://www.geoapps.com/nomime.shtml
Read about quoting: http://www.netmeister.org/news/learn2quote.html
Read about disclaimers: http://www.goldmark.org/jeff/stupid-disclaimers
Dnsbl lookups not working since upgrade to 3.0.0
Posted by marti <ma...@ntlworld.com>.
I am running suse 8.1 just upgraded from 2.64 everything works bar the dnsbl
lookups, not had one positive result since upgrading.
Is there some other perl modules I need to upgrade? Other than the
perl-spamassassin-3.0.0-1.i586.rpm
Martin
Re: Bayes sa-learn change in SA 3.0.1?
Posted by Bill Landry <bi...@pointshare.com>.
Second follow-up:
Please disregard my last post, this ended up being a directory access issue
(yes, user error). The original post under this subject heading is still
valid, though.
=====
debug: rewrite_header: ignoring 5, not From, Subject, or To
debug: bayes: 9607 tie-ing to DB file R/O
/var/amavis/.spamassassin/bayes_toks
debug: bayes: 9607 tie-ing to DB file R/O
/var/amavis/.spamassassin/bayes_seen
debug: bayes: found bayes db version 3
debug: Score set 2 chosen.
debug: Initialising learner
debug: Syncing Bayes and expiring old tokens...
debug: lock: 9607 created /var/amavis/.spamassassin/bayes.mutex
debug: lock: 9607 trying to get lock on /var/amavis/.spamassassin/bayes with
300 timeout
debug: lock: 9607 link to /var/amavis/.spamassassin/bayes.mutex: link ok
debug: bayes: 9607 tie-ing to DB file R/W
/var/amavis/.spamassassin/bayes_toks
debug: bayes: 9607 tie-ing to DB file R/W
/var/amavis/.spamassassin/bayes_seen
debug: bayes: found bayes db version 3
debug: refresh: 9607 refresh /var/amavis/.spamassassin/bayes.mutex
debug: refresh: 9607 refresh /var/amavis/.spamassassin/bayes.mutex
synced Bayes databases from journal in 3 seconds: 3231 unique entries (3534
total entries)
debug: refresh: 9607 refresh /var/amavis/.spamassassin/bayes.mutex
debug: refresh: 9607 refresh /var/amavis/.spamassassin/bayes.mutex
debug: bayes: expiry check keep size, 0.75 * max: 112500
debug: bayes: token count: 157212, final goal reduction size: 44712
debug: bayes: First pass? Current: 1098566094, Last: 1098564710, atime:
59178, count: 0, newdelta: 0, ratio: 0, period: 43200
debug: bayes: Can't use estimation method for expiry, something fishy,
calculating optimal atime delta (first pass)
debug: bayes: expiry max exponent: 9
debug: bayes: atime token reduction
debug: bayes: ======== ===============
debug: bayes: 43200 79188
debug: bayes: 86400 0
debug: bayes: 172800 0
debug: bayes: 345600 0
debug: bayes: 691200 0
debug: bayes: 1382400 0
debug: bayes: 2764800 0
debug: bayes: 5529600 0
debug: bayes: 11059200 0
debug: bayes: 22118400 0
debug: bayes: couldn't find a good delta atime, need more token difference,
skipping expire.
debug: Syncing complete.
debug: bayes: 9607 untie-ing
debug: bayes: 9607 untie-ing db_toks
debug: bayes: 9607 untie-ing db_seen
debug: bayes: files locked, now unlocking lock
debug: unlock: 9607 unlocked /var/amavis/.spamassassin/bayes.mutex
=====
Sorry for the follow-up noise... :-\
Bill
Re: Bayes sa-learn change in SA 3.0.1?
Posted by Bill Landry <bi...@pointshare.com>.
As a follow-up to my own post, since upgrading to SA 3.0.1, I am also
seeing:
=====
debug: open of AWL file failed: lock: 6994 cannot create lockfile
/var/amavis/.spamassassin/auto-whitelist.mutex: Permission denied
debug: bayes: no dbs present, cannot tie DB R/O:
/var/amavis/.spamassassin/bayes_toks
debug: auto-learning failed: lock: 6994 cannot create lockfile
/var/amavis/.spamassassin/bayes.mutex: Permission denied
=====
In local.cf I have "bayes_file_mode 777".
ls -l /var/amavis/.spamassassin/
total 21348
-rw-rw-rw- 1 amavis amavis 10366976 Oct 23 11:41 auto-whitelist
-rw-rw-rw- 1 amavis amavis 6 Oct 23 11:41 auto-whitelist.mutex
-rw-rw-rw- 1 amavis amavis 1476 Oct 23 12:37 bayes.mutex
-rw-rw-rw- 1 amavis amavis 10498048 Oct 23 12:37 bayes_seen
-rw-rw-rw- 1 amavis amavis 5517312 Oct 23 12:37 bayes_toks
However, /usr/bin/sa-learn --dump magic -D looks okay (relevant snippets):
=====
debug: SpamAssassin version 3.0.1
[...]
debug: bayes: 7301 tie-ing to DB file R/O
/var/amavis/.spamassassin/bayes_toks
debug: bayes: 7301 tie-ing to DB file R/O
/var/amavis/.spamassassin/bayes_seen
debug: bayes: found bayes db version 3
debug: Score set 2 chosen.
0.000 0 3 0 non-token data: bayes db version
0.000 0 214770 0 non-token data: nspam
0.000 0 84960 0 non-token data: nham
0.000 0 156069 0 non-token data: ntokens
0.000 0 1098484588 0 non-token data: oldest atime
0.000 0 1098556914 0 non-token data: newest atime
0.000 0 1098559462 0 non-token data: last journal sync
atime
0.000 0 1098560258 0 non-token data: last expiry atime
0.000 0 59178 0 non-token data: last expire atime
delta
0.000 0 0 0 non-token data: last expire
reduction count
debug: bayes: 7301 untie-ing
debug: bayes: 7301 untie-ing db_toks
debug: bayes: 7301 untie-ing db_seen
=====
So what's up with all of this? Looks like SA bayes and awl are both hosed
with this latest release. FYI, I spawn SA via amavisd-new.
Bill
----- Original Message -----
From: "Bill Landry" <bi...@pointshare.com>
To: <us...@SpamAssassin.apache.org>
Sent: Saturday, October 23, 2004 10:56 AM
Subject: Bayes sa-learn change in SA 3.0.1?
> I have "bayes_auto_expire 0" in my local.cf and instead run
> "sa-learn --force-expire" via an hourly cron job. With SA 3.0.0 I was see
> the following results in my hourly cron results e-mails:
> =====
> synced Bayes databases from journal in 3 seconds: 3677 unique entries
(3714
> total entries)
> expired old Bayes database entries in 83 seconds
> 129678 entries kept, 10414 deleted
> token frequency: 1-occurence tokens: 55.30%
> token frequency: less than 8 occurrences: 26.17%
> =====
>
> Since upgrading to SA 3.0.1 last night, I am now only seeing the following
> in my hourly cron results e-mails:
> =====
> synced Bayes databases from journal in 3 seconds: 3268 unique entries
(3478
> total entries)
> =====
>
> I manually ran "sa-learn --force-expire -D" to see what was going on, and
am
> seeing the following lines in my debug output:
> =====
> debug: bayes: Can't use estimation method for expiry, something fishy,
> calculating optimal atime delta (first pass)
> debug: bayes: couldn't find a good delta atime, need more token
difference,
> skipping expire.
> =====
>
> Is this new behavior to be expected now with SA 3.0.1?
>
> Bill
>