Interesting point about the size of the spam db.  I am not in front of
the machine now, but I seem to remember that the spam db ADDs to
itself each rebuild.  Is there a way to have it build to a point and
then remove old data?

thanks


On 5/7/10, Thomas Eckardt <thomas.ecka...@thockar.com> wrote:
> >When assp starts it's using about 780mb of ram.  After running for a
>  >couple of days, it gets up to 1.4gb.  Is that normal?
>
>
> Depending on the size of your hashes and list, this could be normal - and
>  as long there is all running well, there is nothing to change.
>  For example: I run 5 Workerthreads, and I use a DB for all hashes and
>  lists (also the internal). ASSP starts at 300MB and never goes higher than
>  450MB (even on high load systems) - if the rebuild is used, additional 100
>  MB would be needed.
>
>  I think, your resulting Spamdb with 1,774,986 records is too large. A good
>  size would be between 300k and 400k records.
>
>
>  Thomas
>
>
>
>  Von:    K Post <nntp.p...@gmail.com>
>  An:     ASSP development mailing list <assp-test@lists.sourceforge.net>
>
> Datum:  06.05.2010 18:06
>
> Betreff:        Re: [Assp-test] Info logged about sigoff and sigon
>
>
>
>  No databases (yet).  All flat files.
>
>  Seems crazy that it's 3-5 minutes.  Obviously there's something wrong,
>  I just don't know what (and I assume on my end).
>
>  I'm going to try to set this up on a separate machine for testing with
>  the same data to see if that helps.  However, it'll take me a couple
>  of days to have the time.
>
>  When assp starts it's using about 780mb of ram.  After running for a
>  couple of days, it gets up to 1.4gb.  Is that normal?
>
>
>  On Wed, May 5, 2010 at 11:56 AM, Thomas Eckardt
>  <thomas.ecka...@thockar.com> wrote:
>  >> May-01-10 05:43:15 Resulting file 'c:/assp/spamdb.rb.tmp' is 3,908,453
>  > bytes
>  >> May-01-10 05:47:53 Bayesian Pairs: 165,592 in new mail, 1,774,986 now
>  in
>  > list
>  >
>  > The only thing that is done between both lines is
>  >
>  >    &rb_printlog( "\nResulting file '$main::base/spamdb.rb.tmp' is " .
>  > &rb_commify($filesize) . " bytes\n" );
>  >    my $allpairs = scalar keys %main::Spamdb ;
>  >    my $text = ($main::ReplaceOldSpamdb) ? 'new' : 'in new mail';
>  >    &rb_printlog("Bayesian Pairs: " . &rb_commify($addcount) . " $text, "
>  > . &rb_commify($allpairs) . " now in list\n");
>  >    &rb_mlog("Bayesian Pairs: " . &rb_commify($addcount) . " $text, " .
>  > &rb_commify($allpairs) . " now in list");
>  >
>  >
>  > scalar keys %main::Spamdb
>  >
>  > Do you use a DB for that? simply the counting of all records takes 3,5
>  > minutes !?
>  >
>  > Thomas
>  >
>  >
>  >
>  >
>  >
>  > Von:    K Post <nntp.p...@gmail.com>
>  > An:     ASSP development mailing list <assp-test@lists.sourceforge.net>
>  > Datum:  05.05.2010 15:42
>  > Betreff:        Re: [Assp-test] Info logged about sigoff and sigon
>  >
>  >
>  >
>  > (bump)
>  >
>  > Still getting this on and off.
>  > Thanks
>  >
>  > On Sat, May 1, 2010 at 11:31 AM, K Post <nntp.p...@gmail.com> wrote:
>  >> This builds on my previous thread "SMTP not responding every night at
>  >> the same approximte time"
>  >>
>  >>
>  >> Win32, 2.0.2 (1.0.12) -- I haven't updated yet, but have seen similar
>  >> things to this sporatically for a while now.
>  >>
>  >> Here's a clip from the log:
>  >> May-01-10 00:41:33 msg88264-08990 88.199.249.6
>  >> <spamm...@somedomain.com> Connection idle for 180 secs - timeout
>  >> May-01-10 00:41:48 msg88260-10103 88.199.249.6
>  >> <spamm...@somedomain2.com> Connection idle for 180 secs - timeout
>  >> May-01-10 00:42:29 Info: Loop in Worker_2 was not active for 189
>  seconds
>  >> May-01-10 00:42:39 Info: Worker_2 : last sigoff in main,
>  >> c:\assp\assp.pl, 9301, main::sigoffTry, 1, , , at 11 0:3345
>  >> 1272688425.69521 - 9301
>  >> May-01-10 00:43:00 msg88307-04742 [InvalidAddress] 190.166.204.2
>  >> <spamm...@thriddomain.com> invalid address rejected:
>  >> i...@ourdomain.org
>  >> May-01-10 00:43:01 Info: Worker_2 : last sigon in main,
>  >> c:\assp\assp.pl, 9303, main::sigonTry, 1, , , at 11 0:3345
>  >> 1272688425.69543 - 9303
>  >> May-01-10 00:43:01 msg88307-04742 190.166.204.2
>  >> <spamm...@thriddomain.com> Message-Score: added 2 for invalid address
>  >> i...@ourdomain.org, total score for this message is now 2
>  >>
>  >> Lines 9301 and 9301 are:
>  >>    &sigoffTry(__LINE__);
>  >>  and
>  >>   &sigonTry(__LINE__);
>  >>
>  >> I've seen this warning / error before.  It almost always seems to be
>  >> during rebuilds (which this error was).
>  >>
>  >> We've rebuilt since this error, but I know where the slowdown is.
>  >> Look at this excerpt from the rebuild log.
>  >> May-01-10 05:43:15 done - Generating weighted Bayesian tuplets
>  >>
>  >> May-01-10 05:43:15 Resulting file 'c:/assp/spamdb.rb.tmp' is 3,908,453
>  > bytes
>  >> May-01-10 05:47:53 Bayesian Pairs: 165,592 in new mail, 1,774,986 now
>  in
>  > list
>  >>

------------------------------------------------------------------------------

_______________________________________________
Assp-test mailing list
Assp-test@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/assp-test

Reply via email to