I had been toiling with an issue about /var/spool/spamassassin no having
properly updated files -- then I discovered that for some reason, the file
is now being written to /tmp/.spamassassin as user "nobody".
I didn't make any configuration changes for this -- and am a little
puzzled. Perhaps
A new approach to DNSBL might be considered, where there is a peer-to-peer
sharing (authentication, scoring whatever) that mirrors content --
something of that nature, whereby the hackers would basically have to DDos
the entire internet to prevent its use. Not sure how such a framework
could
I'm also curious about how this virus/worm operates. I'm presuming it's
infiltrating Outlook address books, but I could be wrong - since my server
is being hit with 100's of these messages, and that indicates possibly some
other source of data (for addresses).
I'm successfully blocking these w
This new virus appears to generate many (random?) subjects, so it's getting
difficult to narrow down.
Has anyone filters for Spamassassin that will correctly identify this
virus? I'd like to score this one high so they are rejected (via
spamass-milter)... it's been a huge problem all day.
The
Same here, on FreeBSD-4.9-Prerelease.
Gives an error, but seems to work alright.
At 10:21 AM 9/19/2003, Christopher X. Candreva wrote:
Under rc5, spamd gives this warning on start-up:
Use of uninitialized value in scalar assignment at
/usr/local/lib/perl5/site_perl/5.6.1/Mail/SpamAssassin/U
This new virus appears to generate many (random?) subjects, so it's getting
difficult to narrow down.
Has anyone filters for Spamassassin that will correctly identify this
virus? I'd like to score this one high so they are rejected (via
spamass-milter)... it's been a huge problem all day.
The
I used the switch -D and here's the output -- it looks like something
might be off. Since I've already updated the
/var/spool/spamassassin/bayes files... and the timestamps were just updated
now with the SA import:
-rw-r--r-- 1 root wheel 1633 Aug 29 11:30 bayes_msgcount
-rw--- 1
Van Dinter wrote:
On Sat, Sep 06, 2003 at 03:06:29PM -0400, Forrest Aldrich wrote:
> And I'm not certain that these files are actually being used by SA. The
> bayse_msgcount, for example, hasn't been update for days. The update on
msgcount won't be used in 2.60. in fact, when
I followed the directions you gave, and now have in /var/spool/spamassassin:
-rw-r--r-- 1 root wheel 1633 Aug 29 11:30 bayes_msgcount
-rw--- 1 root wheel 630784 Sep 5 22:32 bayes_seen
-rw--- 1 root wheel 2809856 Sep 5 22:32 bayes_toks
-rw-r--r-- 1 root wheel 630784 Sep
error.
Seems something is really wrong with this command.
At 05:27 PM 9/5/2003, Malte S. Stretz wrote:
On Friday 05 September 2003 23:01 CET Forrest Aldrich wrote:
> I actually did try that, and I get this error:
>
> # sa-learn --import --dbpath /var/spool/spamassassi
I actually did try that, and I get this error:
# sa-learn --import --dbpath /var/spool/spamassassin
bayes upgrade_old_dbm_files: unable to find bayes_toks and bayes_seen, stopping
At 04:33 PM 9/5/2003, Malte S. Stretz wrote:
On Friday 05 September 2003 22:08 CET Forrest Aldrich wrote:
>
I installed RC3 on FreeBSD-4.x and ran sa-learn --import on the bayes_*
files in /var/spool/spamassassin, but I don't believe this worked
correctly, as the files don't seem to get updated from thereon:
-rw-r--r-- 1 root wheel 1633 Aug 29 11:30 bayes_msgcount
-rw--- 1 root wheel 630
OT question: I've recently been seeing these types of entries in my mail
log (sendmail 8.12.9+SA-2.55+SA-Milter) which I can't explain. It would
seem that spamd is getting something confused when analyzing inbound
messages from the spamassassin-talk list -- this isn't happening in any
other
Curious if anyone has tried this. It's probably wide open to a DoS...
Adding a routine to the Spam collecting process -- either through milter,
SpamAssassin itself, or Mailscanner, such that a counter is maintained for
each Class C address, incremented for each message that SA detects coming
t
After reading the FAQ, I decided it would be appropriate to ask about this.
Regardless of MTA-specific criteria, I wondered if there was a way (or an
effort to create) to integrate SpamAssasin at the MTA level (Sendmail,
Postfix, Exim, Qmail -- whatever). The goal being to further insulate
re
15 matches
Mail list logo
