You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Justin Mason <jm...@jmason.org> on 2004/07/30 00:40:20 UTC

Re: Pyzor issue with Amavisd-New 2.0 & SA 3.0.0-pre3

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Bill Landry writes:
> I've haven't seen any responses to the issue yet.  Is this something that
> the devs want to resolve before the SA 3.0 release or is Pyzor support being
> phased out of SA?

if you want the devs to respond, opening a bug on the bugzilla would be
better -- posting on -users is likely to get lost in the noise (it's
a very busy list nowadays ;)

anyway, yep, open a bug for this one, it sounds like it needs some
attention.

- --j.

> Bill
> ----- Original Message ----- 
> From: "Bill Landry" <bi...@pointshare.com>
> To: <sp...@incubator.apache.org>;
> <am...@lists.sourceforge.net>
> Sent: Saturday, July 24, 2004 10:54 AM
> Subject: Re: Pyzor issue with Amavisd-New 2.0 & SA 3.0.0-pre3
> 
> > My original post is below.  Here is some more detail on this issue.  When
> > running Amavisd-New 2.0 and SA 2.63, Pyzor, Razor, and DCC all run fine.
> > However, after upgrading to SA 3.0-pre3, Pyzor times out on all queries,
> but
> > Razor and DDC still run fine:
> >
> > Amavisd-New 2.0 & SA 2.63:
> > ==========
> > debug: Razor2 is available
> > debug: Pyzor is available: /usr/local/bin/pyzor
> > debug: entering helper-app run mode
> > debug: Pyzor: got response: 66.250.40.33:24441  (200, 'OK')     0       0
> > debug: leaving helper-app run mode
> > debug: DCCifd is available: /var/dcc/dccifd
> > debug: entering helper-app run mode
> > debug: DCCifd: got response: X-DCC-dmv.com-Metrics: gw2.pointshare.com
> 1181;
> > Body=1 Fuz1=1 Fuz2=1
> > debug: leaving helper-app run mode
> > ==========
> >
> > Amavisd-New 2.0 & SA 3.0-pre3:
> > ==========
> > debug: Razor2 is available
> > debug: Pyzor is available: /usr/local/bin/pyzor
> > debug: entering helper-app run mode
> > debug: setuid: helper proc 9392: ruid=3000 euid=3000
> > debug: leaving helper-app run mode
> > debug: Pyzor -> check timed out after 10 secs.
> > debug: DCCifd is available: /var/dcc/dccifd
> > debug: entering helper-app run mode
> > debug: DCCifd: got response: X-DCC-sonic.net-Metrics: mgw1.pointshare.com
> > 1117; Body=1 Fuz1=1 Fuz2=1
> > debug: leaving helper-app run mode
> > ==========
> >
> > It appears that the only difference between SA 2.63 and SA 3.0-pre3 is the
> > addition of the "setuid" line:
> >
> >      debug: setuid: helper proc 9392: ruid=3000 euid=3000
> >
> > UID 3000 & GID 3000 are owned by Amavisd-New.  Was this "setuid" process
> an
> > intentional change, as it would seem to restrict Amavisd-New's ability to
> > access the Python/Pyzor applications as they are owned by root.  I have
> > attached an strace snippet that shows that with SA 3.0-pre3, when
> attempting
> > to call Python and Pyzor, the process experiences lots of "Bad file
> > descriptor" issues, which are not present with SA 2.63.
> >
> > Hopefully one of the devs will respond with some feedback on this.  Should
> I
> > log this as a bug, or is there a workaround for this issue that I'm not
> > seeing.
> >
> > Regards,
> >
> > Bill
> >
> > PS, wanted to thank William Stearns for the nice "apptrace" tool
> (available
> > at www.stearns.org), it helped a lot to be able to capture what was
> > happening with this issue.
> >
> > ----- Original Message ----- 
> > From: "Bill Landry" <bi...@pointshare.com>
> > To: <sp...@incubator.apache.org>;
> > <am...@lists.sourceforge.net>
> > Sent: Friday, July 23, 2004 3:16 PM
> > Subject: Pyzor issue with Amavisd-New 2.0 & SA 3.0.0-pre3
> >
> >
> > > I am using amavisd-new 2.0 to call spamassassin and when I start amavisd
> > in
> > > debug-sa mode, all helper-apps test fine, including pyzor:
> > > ==========
> > > debug: Pyzor is available: /usr/bin/pyzor
> > > debug: entering helper-app run mode
> > > debug: setuid: helper proc 20615: ruid=3000 euid=3000
> > > debug: Pyzor: got response: 66.250.40.33:24441  (200, 'OK')     0
> 0
> > > debug: leaving helper-app run mode
> > > ==========
> > >
> > > And when piping a test message into SA, all is fine, as well:
> > >  spamassassin -dtD < test.msg
> > > ==========
> > > debug: Pyzor is available: /usr/bin/pyzor
> > > debug: entering helper-app run mode
> > > debug: setuid: helper proc 2513: ruid=0 euid=0
> > > debug: Pyzor: got response: 66.250.40.33:24441  (200, 'OK')     47
> 0
> > > debug: leaving helper-app run mode
> > > ==========
> > >
> > > However, once amavisd-new has fully started and mail is received, pyzor
> > > times out on every message:
> > > ==========
> > > debug: Pyzor is available: /usr/bin/pyzor
> > > debug: entering helper-app run mode
> > > debug: setuid: helper proc 2602: ruid=3000 euid=3000
> > > debug: leaving helper-app run mode
> > > debug: Pyzor -> check timed out after 10 secs.
> > > ==========
> > >
> > > Pyzor works fine with SA 2.61, and all other helper-apps (Razor & DCC)
> > work
> > > fine with 2.61 and 3.0, and I can still ping the pyzor server - "pyzor
> > ping"
> > > comes back with:
> > > ==========
> > > 66.250.40.33:24441      (200, 'OK')
> > > ==========
> > >
> > > I am cross-posting this to both the SA and Amavis lists since I am not
> > sure
> > > where the issues is.  Any feedback or suggestions are most welcome.
> > >
> > > Regards,
> > >
> > > Bill
> > >
> >
> >
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Exmh CVS

iD8DBQFBCXzUQTcbUG5Y7woRAiKGAJ4/L5malkbKEKnV9ihcC0/x3lms/gCgvHFW
inm2xTGG3DaDuMjUV60VC5g=
=kvPG
-----END PGP SIGNATURE-----


Re: Pyzor issue with Amavisd-New 2.0 & SA 3.0.0-pre3

Posted by Bill Landry <bi...@pointshare.com>.
----- Original Message ----- 
From: "Justin Mason" <jm...@jmason.org>

> > I've haven't seen any responses to the issue yet.  Is this something
that
> > the devs want to resolve before the SA 3.0 release or is Pyzor support
being
> > phased out of SA?
>
> if you want the devs to respond, opening a bug on the bugzilla would be
> better -- posting on -users is likely to get lost in the noise (it's
> a very busy list nowadays ;)
>
> anyway, yep, open a bug for this one, it sounds like it needs some
> attention.

Done: Bug 3649

Thanks,

Bill