You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by George Georgalis <ge...@galis.org> on 2005/09/07 17:08:24 UTC
not using rbls?
it would seem the following rule is not being used...
header __RCVD_IN_SBL_XBL eval:check_rbl('sblxbl', 'sbl-xbl.spamhaus.org.')
describe __RCVD_IN_SBL_XBL Received via a relay in Spamhaus SBL+XBL
tflags __RCVD_IN_SBL_XBL net
I do have Mail::SpamAssassin::Plugin::URIDNSBL but maybe I need another plugin?
What's missing?
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by George Georgalis <ge...@galis.org>.
On Wed, Sep 07, 2005 at 05:31:13PM -0400, Matt Kettler wrote:
>George Georgalis wrote:
>
>>
>> Well since my MTA doesn't have an IP put in the header (it gets
>> mail from stdin -- and SA runs from stdin forwarding the exit code
>> to the MTA which then accept or rejects to the remote relay), I
>> think I'll have to settle with incomplete functionality.
>
>Your MTA doesn't need to put it's IP in the received: header. Most MTA's don't.
>Mine doesn't. The insertion can be a hostname as long as that's resolvable via DNS.
>
>Looking at this, your MTA is inserting a perfectly usable Received header:
>
>2005-09-07 13:49:10.986551500 debug: received-header: parsed as [
>ip=83.x2.166.71 rdns=eac71.neoplus.adsl.tpnet.pl
>helo=eac71.neoplus.adsl.tpnet.pl by=sta.galis.org ident= envfrom= intl=0 id= auth= ]
>
>But whatever sta.galis.org resolves to isn't trusted, that's the problem.
>
>Put the IP that your spamassassin box will get if it does a DNS lookup of
>"sta.galis.org" into your trusted_networks.
>
>i.e: I resolve sta.galis.org as 66.250.170.210, so I'd put 66.250.170.210/32 in
>my trusted networks.
I tried using both the NAT subnet the MTA is on and the external subnet
trusted_networks 127.0.0.1 66.250.170.208/28 192.168.80.0/24
with no change
2005-09-07 20:56:23.954657500 debug: received-header: parsed as [ ip=83.x2.166.71 rdns=eac71.neoplus.adsl.tpnet_pl helo=eac71.neoplus.adsl.tpnet_pl by=sta.galis.org ident= envfrom= intl=0 id= auth= ]
2005-09-07 20:56:23.954666500 debug: received-header: relay 83.x2.166.71 trusted? no internal? no
2005-09-07 20:56:23.954671500 debug: metadata: X-Spam-Relays-Trusted:
2005-09-07 20:56:23.954675500 debug: metadata: X-Spam-Relays-Untrusted: [ ip=83.x2.166.71 rdns=eac71.neoplus.adsl.tpnet_pl helo=eac71.neoplus.adsl.tpnet_pl by=sta.galis.org ident= envfrom= intl=0 id= auth= ]
the local _resolver_ does work correctly, eg not per /etc/hosts
PING sta.galis.org (192.168.80.50): 56 data bytes
64 bytes from 192.168.80.50: icmp_seq=0 ttl=64 time=0.1 ms
64 bytes from 192.168.80.50: icmp_seq=1 ttl=64 time=0.1 ms
so I'm not sure why X-Spam-Relays-Trusted is empty.
>Fix that, as I told you before, and quit worrying about stdin.
Sorry, I think I said that twice because I was following up on two
sub threads, I did try using trusted_networks when I first read
that message. With same results as per above, not sure what the
problem is.
>> Shouldn't the default be, if nothing is trusted, test the first
>> relay?
>
>No, because that's not safe.
>
>If there's no trusted Received: headers, SA cannot make any safe assumptions
>about what IPs are or are not a part of your network.
My thinking is (at least) the top one must be reliable, if you cannot
trust your top MTA you could use 127.0.0.1 as trusted or maybe not
even use rbl, but I have no reason to debate this.
>At a very fundamental level it is complete garbage input. There MUST be a
>Received: header that can be trusted. If there are none, the message shouldn't
>be there.
Well, got to agree there.
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by Matt Kettler <mk...@evi-inc.com>.
George Georgalis wrote:
>
> Well since my MTA doesn't have an IP put in the header (it gets
> mail from stdin -- and SA runs from stdin forwarding the exit code
> to the MTA which then accept or rejects to the remote relay), I
> think I'll have to settle with incomplete functionality.
Your MTA doesn't need to put it's IP in the received: header. Most MTA's don't.
Mine doesn't. The insertion can be a hostname as long as that's resolvable via DNS.
Looking at this, your MTA is inserting a perfectly usable Received header:
2005-09-07 13:49:10.986551500 debug: received-header: parsed as [
ip=83.x2.166.71 rdns=eac71.neoplus.adsl.tpnet.pl
helo=eac71.neoplus.adsl.tpnet.pl by=sta.galis.org ident= envfrom= intl=0 id= auth= ]
But whatever sta.galis.org resolves to isn't trusted, that's the problem.
Put the IP that your spamassassin box will get if it does a DNS lookup of
"sta.galis.org" into your trusted_networks.
i.e: I resolve sta.galis.org as 66.250.170.210, so I'd put 66.250.170.210/32 in
my trusted networks.
Fix that, as I told you before, and quit worrying about stdin.
>
> Shouldn't the default be, if nothing is trusted, test the first
> relay?
No, because that's not safe.
If there's no trusted Received: headers, SA cannot make any safe assumptions
about what IPs are or are not a part of your network.
At a very fundamental level it is complete garbage input. There MUST be a
Received: header that can be trusted. If there are none, the message shouldn't
be there.
r.
Re: not using rbls?
Posted by George Georgalis <ge...@galis.org>.
On Wed, Sep 07, 2005 at 03:45:56PM -0400, Matt Kettler wrote:
>George Georgalis wrote:
>> In my setup, trusted relays arn't tested with SA, they go straight
>> to the queue. Untrusted networks must negociate SA in SMTP. I've
>> visited this configuration issue before, and the simple solution
>> has been
>
>NOOOOOOOOOOOO. You have the wrong idea.
>
>Trusted here does NOT mean trusted to never relay spam, it means trusted to not
>forge headers.
>
>You MUST trust your own mailserver that receives mail from the internet. MUST!
>Even if it accepts spam, that's fine. But it has to be trusted. For SA to work
>properly you have to trust it.
Well since my MTA doesn't have an IP put in the header (it gets
mail from stdin -- and SA runs from stdin forwarding the exit code
to the MTA which then accept or rejects to the remote relay), I
think I'll have to settle with incomplete functionality. Which is
not so bad because I already do pass/fail per network before SA
starts, this particular xbl test is missing from my front-end, for
now.
>If set up properly this will NOT cause Internet mail to hit ALL_TRUSTED, that's
>the result of the opposite extreme of the same problem.. too much trust. Right
>now you've got too little trust.
Well, I'm not 100% clear on the what is SA proper in this
situation.
>Doing what you have done breaks a LOT of SA's rules, including
>whitelist_from_rcvd, all dialup RBLs, etc. Without any trusted relays, SA
>doesn't know where the mail entered your network, so it won't run any tests that
>rely on recognizing the boundaries between your network and the Internet.
Shouldn't the default be, if nothing is trusted, test the first
relay?
>Set your trusted_networks properly, and remove your hack-fix of reducing
>ALL_TRUSTED's score. Set trusted_networks to the IPs of all mail relays you
>control and trust to not forge headers.
Per above (no IP), I don't think that will help.
>I suspect all your problems started because your outside mail relay is NATed,
>and you hit the "over trust" bug where ALL_TRUSTED matches mail from the outside.
yes.
>To fix this, you over-reacted and created an under-trusting configuration.
>That's just as broken.
>
>Read this wiki article for the lengthy details:
>http://wiki.apache.org/spamassassin/TrustPath
I see three options, me using another MTA; SA is patched to
assume the first received is by a trusted network (eg go ahead and
do rbl tests on it if the user didn't define trusted networks);
or skip_rbl_checks and do tests prior to SA.
Such is life, I'm open to alternatives. Thanks for your time.
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by Matt Kettler <mk...@evi-inc.com>.
George Georgalis wrote:
> In my setup, trusted relays arn't tested with SA, they go straight
> to the queue. Untrusted networks must negociate SA in SMTP. I've
> visited this configuration issue before, and the simple solution
> has been
NOOOOOOOOOOOO. You have the wrong idea.
Trusted here does NOT mean trusted to never relay spam, it means trusted to not
forge headers.
You MUST trust your own mailserver that receives mail from the internet. MUST!
Even if it accepts spam, that's fine. But it has to be trusted. For SA to work
properly you have to trust it.
If set up properly this will NOT cause Internet mail to hit ALL_TRUSTED, that's
the result of the opposite extreme of the same problem.. too much trust. Right
now you've got too little trust.
Doing what you have done breaks a LOT of SA's rules, including
whitelist_from_rcvd, all dialup RBLs, etc. Without any trusted relays, SA
doesn't know where the mail entered your network, so it won't run any tests that
rely on recognizing the boundaries between your network and the Internet.
Set your trusted_networks properly, and remove your hack-fix of reducing
ALL_TRUSTED's score. Set trusted_networks to the IPs of all mail relays you
control and trust to not forge headers.
I suspect all your problems started because your outside mail relay is NATed,
and you hit the "over trust" bug where ALL_TRUSTED matches mail from the outside.
To fix this, you over-reacted and created an under-trusting configuration.
That's just as broken.
Read this wiki article for the lengthy details:
http://wiki.apache.org/spamassassin/TrustPath
Re: not using rbls?
Posted by George Georgalis <ge...@galis.org>.
On Wed, Sep 07, 2005 at 02:49:50PM -0400, Matt Kettler wrote:
>George Georgalis wrote:
>> 2005-09-07 13:49:10.975816500 logmsg: checking message <10...@eac71.neoplus.adsl.tpnet.pl> for geo:1002.
>> 2005-09-07 13:49:10.986551500 debug: received-header: parsed as [ ip=83.x2.166.71 rdns=eac71.neoplus.adsl.tpnet.pl helo=eac71.neoplus.adsl.tpnet.pl by=sta.galis.org ident= envfrom= intl=0 id= auth= ]
>> 2005-09-07 13:49:10.989047500 debug: received-header: relay 83.x2.166.71 trusted? no internal? no
>> 2005-09-07 13:49:10.989055500 debug: metadata: X-Spam-Relays-Trusted:
>
>Danger!!! There are no trusted relays. That should *never* happen!
>
>Fix your trusted_networks such that the IP of sta.galis.org is trusted.
In my setup, trusted relays arn't tested with SA, they go straight
to the queue. Untrusted networks must negociate SA in SMTP. I've
visited this configuration issue before, and the simple solution
has been
trusted_networks 127.0.0.1
score ALL_TRUSTED -0.001
I just tried both using the sta.galis.org name and ip, no go,
presumably because the my MTA does not add the local IP for
received mail, only the name
Received: from eac71.neoplus.adsl.tpnet_pl (83.x2.166.71)
by sta.galis.org with SMTP; 7 Sep 2005 10:00:34 -0000
(my underscore and x) do you think this has anything to do
with my xbl problem?
>In this case, the lack of trust is the cause of your problem. The IP in question
>is listed in XBL, but XBL will only be tested against the host delivering mail
>to the first trusted server.
>
>Since there are no trusted servers, there will never be a match against XBL. Ever.
well that's no fun. I do have an environmental ${TCPREMOTEIP}, it
sounds like I'll need to do the query in house or patch my MTA, to
list ip addresses with hostnames (won't happen).
Would it be reasonable for SA to test the FIRST received ip if no
trusted are found? (and optionally additional ones) The requiring
a "trusted_networks" to function seems to restrict the ways SA can
be deployed.
In any event a pre-SA XBL test will probably go a long ways to
lighten the system load. Thanks.
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by Matt Kettler <mk...@evi-inc.com>.
George Georgalis wrote:
> 2005-09-07 13:49:10.975816500 logmsg: checking message <10...@eac71.neoplus.adsl.tpnet.pl> for geo:1002.
> 2005-09-07 13:49:10.986551500 debug: received-header: parsed as [ ip=83.x2.166.71 rdns=eac71.neoplus.adsl.tpnet.pl helo=eac71.neoplus.adsl.tpnet.pl by=sta.galis.org ident= envfrom= intl=0 id= auth= ]
> 2005-09-07 13:49:10.989047500 debug: received-header: relay 83.x2.166.71 trusted? no internal? no
> 2005-09-07 13:49:10.989055500 debug: metadata: X-Spam-Relays-Trusted:
Danger!!! There are no trusted relays. That should *never* happen!
Fix your trusted_networks such that the IP of sta.galis.org is trusted.
In this case, the lack of trust is the cause of your problem. The IP in question
is listed in XBL, but XBL will only be tested against the host delivering mail
to the first trusted server.
Since there are no trusted servers, there will never be a match against XBL. Ever.
Re: not using rbls?
Posted by George Georgalis <ge...@galis.org>.
On Wed, Sep 07, 2005 at 02:15:09PM -0400, Fred wrote:
>George Georgalis wrote:
>> Argument "REPORT" isn't numeric in
>> subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223,
>> <GEN3> line 58. 2005-09-07 13:49:11.026387500 Argument "REPORT" isn't
>> numeric in subroutine entry at
>> /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 58.
>> 2005-09-07 13:49:11.030191500 Argument "REPORT" isn't numeric in
>> subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223,
>> <GEN3> line 58.
>
>>>From the looks of this, something went wrong with your Net::DNS, you might
>consider reinstalling it.
>
I saw those REPORT lines, but the thing is, I installed net::dns
from just before I ran that test...
Writing /usr/local/lib/perl/5.6.1/auto/Net/DNS/.packlist
Appending installation info to /usr/local/lib/perl/5.6.1/perllocal.pod
/usr/bin/make install UNINST=1 -- OK
oh btw I use force install because the online test failed...
t/08-online................NOK 73
# Failed test (t/08-online.t at line 158)
# Failed test (t/08-online.t at line 172)
156 # $res->debug(1);
157 my $socket=$res->bgsend('a.t.net-dns.org','A');
158 ok(ref($socket)=~/^IO::Socket::INET(6?)$/,"Socket returned");
159 my $loop=0;
160 # burn a little CPU to get the socket ready.
161 # I could off course used microsleep or something.
162 while ($loop<200000){
163 $loop++;
164 }
165 $loop=0;
166
167 while ($loop<6){
168 last if $res->bgisready($socket);
169 sleep(1); # If burning CPU above was not sufficient.
170 $loop++;
171 }
172 ok ($res->bgisready($socket),"Socket is ready");
...again, surbl does work.... and I'm pretty sure I'm using _that_ net::dns
because it's the only one
# ls /usr/share/perl5/Net /usr/share/perl/5.6.1/Net
/usr/share/perl/5.6.1/Net:
Ping.pm hostent.pm netent.pm protoent.pm servent.pm
/usr/share/perl5/Net:
Cmd.pm Domain.pm FTP.pm HTTP.pm NNTP.pm POP3.pm Time.pm
Config.pm FTP HTTP HTTPS.pm Netrc.pm SMTP.pm libnetFAQ.pod
could something else be causing the REPORT error/warning?
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by George Georgalis <ge...@galis.org>.
On Wed, Sep 07, 2005 at 03:47:24PM -0400, George Georgalis wrote:
>On Wed, Sep 07, 2005 at 03:24:50PM -0400, Matt Kettler wrote:
>>George Georgalis wrote:
>>
>>> grep -r REPORT /etc/spamassassin/ /usr/share/spamassassin/
>>> returned nothing that wasn't attached to other characters.
>>
>>Hmm.. what about greping /etc/resolv.conf?
>
>nothing.
>
>>Better yet, let's look at the code surrounding line 233 of INET.pm
>> head -n 235 /usr/share/perl/5.6.1/IO/Socket/INET.pm line |tail -n 5
>
> head -n 235 /usr/share/perl/5.6.1/IO/Socket/INET.pm |tail -n 5
>
>}
>
>sub sockport {
> @_ == 1 or croak 'usage: $sock->sockport()';
> my($sock) = @_;
>
>
>>That will at least help us narrow in on what parameter is text instead of
>>numeric, which might help us figure out what the problem is.
Well a grep Argument of my logs brings up some other entries,
2005-09-07 20:49:40.686200500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 117.
2005-09-07 20:49:40.703091500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 117.
2005-09-07 20:49:40.707188500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 117.
2005-09-07 20:49:40.724947500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 117.
2005-09-07 20:53:58.597992500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 80.
2005-09-07 20:53:58.602607500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 80.
2005-09-07 20:53:58.671268500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 80.
2005-09-07 20:53:59.208889500 Argument "ns1.nlnetlabs.nl." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 80.
2005-09-07 20:53:59.212616500 Argument "ns2.berdom.net." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 80.
2005-09-07 20:53:59.216305500 Argument "ns1.berdom.net." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 80.
2005-09-07 20:56:23.962210500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 58.
2005-09-07 20:56:23.978231500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 58.
2005-09-07 20:56:23.982179500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 58.
2005-09-07 20:56:23.999445500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 58.
2005-09-07 21:01:37.253175500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 203.
2005-09-07 21:01:37.253182500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 203.
2005-09-07 21:01:37.302893500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 203.
2005-09-07 21:01:38.623757500 Argument "ns2.info.nl." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 203.
2005-09-07 21:01:38.627453500 Argument "ns1.info.nl." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN18> line 203.
2005-09-07 21:03:44.712295500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN40> line 33.
2005-09-07 21:03:44.712306500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN40> line 33.
2005-09-07 21:03:44.728381500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN40> line 33.
2005-09-07 21:03:44.976395500 Argument "ns1.logingcramp.com." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN40> line 33.
2005-09-07 21:03:44.980485500 Argument "ns2.logingcramp.com." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN40> line 33.
2005-09-07 21:16:24.919024500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN33> line 34.
2005-09-07 21:16:24.923417500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN33> line 34.
2005-09-07 21:16:24.954193500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN33> line 34.
2005-09-07 21:16:25.142212500 Argument "ns1.hushdns.com." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN33> line 34.
2005-09-07 21:16:25.146072500 Argument "ns2.hushdns.com." isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN33> line 34.
2005-09-07 21:21:32.709549500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN44> line 24.
2005-09-07 21:21:33.130575500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN64> line 20.
2005-09-07 21:21:33.327624500 Argument "CHECK" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN51> line 100.
I am able to resolve those domains to IPs.
This time I've managed to include the whole subroutine...
/usr/share/perl/5.6.1/IO/Socket/INET.pm
233 sub sockport {
234 @_ == 1 or croak 'usage: $sock->sockport()';
235 my($sock) = @_;
236 my $name = $sock->sockname;
237 $name ? (sockaddr_in($name))[0] : undef;
238 }
If I can provide any other info, let me know.
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by George Georgalis <ge...@galis.org>.
On Wed, Sep 07, 2005 at 03:24:50PM -0400, Matt Kettler wrote:
>George Georgalis wrote:
>
>> grep -r REPORT /etc/spamassassin/ /usr/share/spamassassin/
>> returned nothing that wasn't attached to other characters.
>
>Hmm.. what about greping /etc/resolv.conf?
nothing.
>Better yet, let's look at the code surrounding line 233 of INET.pm
> head -n 235 /usr/share/perl/5.6.1/IO/Socket/INET.pm line |tail -n 5
head -n 235 /usr/share/perl/5.6.1/IO/Socket/INET.pm |tail -n 5
}
sub sockport {
@_ == 1 or croak 'usage: $sock->sockport()';
my($sock) = @_;
>That will at least help us narrow in on what parameter is text instead of
>numeric, which might help us figure out what the problem is.
could this be a result of no numeric ip from my MTA :)
That's a can't/won't fix because my mta doesn't have an IP only a
name, it gets it's messages via stdin and is invoked by a variety
of sources. See prior email by date in thread re trusted_networks.
Seems like a good reason to allow for NULL trusted networks or to
just use the first Received ip if no trusted networks match. This
seems to cause the problem, but I'm not sure how I would have
anticipated it, from a user's perspective.
>> but this is interesting,
>> ...
>
>That doesn't look at all interesting to me. None of that code has anything to do
>with the network code, so none of that should ever approach INET.pm. That's all
>stuff related to the markup of the message.
k
>>Looks like a bug in my version.
>> # spamassassin -V
>> SpamAssassin version 3.0.4
>> running on Perl version 5.6.1
>
>Highly unlikely to be a bug in your version. a LOT of people run SA 3.0.4 on
>perl 5.6.1, and that particular message is very new to me.
k ;)
I'm open to suggestions.
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by Matt Kettler <mk...@evi-inc.com>.
George Georgalis wrote:
> grep -r REPORT /etc/spamassassin/ /usr/share/spamassassin/
> returned nothing that wasn't attached to other characters.
Hmm.. what about greping /etc/resolv.conf?
Better yet, let's look at the code surrounding line 233 of INET.pm
head -n 235 /usr/share/perl/5.6.1/IO/Socket/INET.pm line |tail -n 5
That will at least help us narrow in on what parameter is text instead of
numeric, which might help us figure out what the problem is.
>
> but this is interesting,
>
> # grep -r REPORT /usr/share/perl5 /usr/share/perl /usr/local/lib/perl/5.6.1/
> /usr/share/perl5/Net/FTP.pm:=head1 REPORTING BUGS
> /usr/share/perl5/HTML/TreeBuilder.pm:=head1 BUG REPORTS
> /usr/share/perl5/Mail/SpamAssassin/Conf.pm: $self->{headers_spam}->{"Report"} = "_REPORT_";
> /usr/share/perl5/Mail/SpamAssassin/Conf.pm: _REPORT_ terse report of tests hit (for header reports)
> /usr/share/perl5/Mail/SpamAssassin/PerMsgStatus.pm: REPORT => sub {
> /usr/share/perl5/Mail/SpamAssassin/PerMsgStatus.pm: return "\n" . ($self->{tag_data}->{REPORT} || "");
> /usr/share/perl5/Mail/SpamAssassin/PerMsgStatus.pm: $self->{tag_data}->{REPORT} .= sprintf ("* %s %s %s%s\n%s",
That doesn't look at all interesting to me. None of that code has anything to do
with the network code, so none of that should ever approach INET.pm. That's all
stuff related to the markup of the message.
>Looks like a bug in my version.
> # spamassassin -V
> SpamAssassin version 3.0.4
> running on Perl version 5.6.1
Highly unlikely to be a bug in your version. a LOT of people run SA 3.0.4 on
perl 5.6.1, and that particular message is very new to me.
Re: not using rbls?
Posted by George Georgalis <ge...@galis.org>.
On Wed, Sep 07, 2005 at 02:47:04PM -0400, Matt Kettler wrote:
>Fred wrote:
>> George Georgalis wrote:
>>
>>>Argument "REPORT" isn't numeric in
>>>subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223,
>>><GEN3> line 58. 2005-09-07 13:49:11.026387500 Argument "REPORT" isn't
>>>numeric in subroutine entry at
>>>/usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 58.
>>>2005-09-07 13:49:11.030191500 Argument "REPORT" isn't numeric in
>>>subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223,
>>><GEN3> line 58.
>>
>>
>>>>From the looks of this, something went wrong with your Net::DNS, you might
>> consider reinstalling it.
>>
>> Frederic Tarasevicius
>>
>
>That's possible, but more likely George has absolute garbage somewhere in his
>config file. Most of the "argument isn't numeric" is the result of blatantly
>bogus input.
>
>In this case, it looks like the text "REPORT" is somewhere that a number belongs
>in the config file.
grep -r REPORT /etc/spamassassin/ /usr/share/spamassassin/
returned nothing that wasn't attached to other characters.
but this is interesting,
# grep -r REPORT /usr/share/perl5 /usr/share/perl /usr/local/lib/perl/5.6.1/
/usr/share/perl5/Net/FTP.pm:=head1 REPORTING BUGS
/usr/share/perl5/HTML/TreeBuilder.pm:=head1 BUG REPORTS
/usr/share/perl5/Mail/SpamAssassin/Conf.pm: $self->{headers_spam}->{"Report"} = "_REPORT_";
/usr/share/perl5/Mail/SpamAssassin/Conf.pm: _REPORT_ terse report of tests hit (for header reports)
/usr/share/perl5/Mail/SpamAssassin/PerMsgStatus.pm: REPORT => sub {
/usr/share/perl5/Mail/SpamAssassin/PerMsgStatus.pm: return "\n" . ($self->{tag_data}->{REPORT} || "");
/usr/share/perl5/Mail/SpamAssassin/PerMsgStatus.pm: $self->{tag_data}->{REPORT} .= sprintf ("* %s %s %s%s\n%s",
Unfortunatly, I'm no perl programmer nor do I know the inner
workings of SA. Looks like a bug in my version.
# spamassassin -V
SpamAssassin version 3.0.4
running on Perl version 5.6.1
Anything I can do?
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by Matt Kettler <mk...@evi-inc.com>.
Fred wrote:
> George Georgalis wrote:
>
>>Argument "REPORT" isn't numeric in
>>subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223,
>><GEN3> line 58. 2005-09-07 13:49:11.026387500 Argument "REPORT" isn't
>>numeric in subroutine entry at
>>/usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 58.
>>2005-09-07 13:49:11.030191500 Argument "REPORT" isn't numeric in
>>subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223,
>><GEN3> line 58.
>
>
>>>From the looks of this, something went wrong with your Net::DNS, you might
> consider reinstalling it.
>
> Frederic Tarasevicius
>
That's possible, but more likely George has absolute garbage somewhere in his
config file. Most of the "argument isn't numeric" is the result of blatantly
bogus input.
In this case, it looks like the text "REPORT" is somewhere that a number belongs
in the config file.
I'd suggest a check like this:
grep REPORT /etc/mail/spamassassin/*.cf
Re: not using rbls?
Posted by Fred <sp...@freddyt.com>.
George Georgalis wrote:
> Argument "REPORT" isn't numeric in
> subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223,
> <GEN3> line 58. 2005-09-07 13:49:11.026387500 Argument "REPORT" isn't
> numeric in subroutine entry at
> /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 58.
> 2005-09-07 13:49:11.030191500 Argument "REPORT" isn't numeric in
> subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223,
> <GEN3> line 58.
>From the looks of this, something went wrong with your Net::DNS, you might
consider reinstalling it.
Frederic Tarasevicius
Re: not using rbls?
Posted by George Georgalis <ge...@galis.org>.
On Wed, Sep 07, 2005 at 11:35:53AM -0400, Matt Kettler wrote:
>George Georgalis wrote:
>> it would seem the following rule is not being used...
>>
>> header __RCVD_IN_SBL_XBL eval:check_rbl('sblxbl', 'sbl-xbl.spamhaus.org.')
>> describe __RCVD_IN_SBL_XBL Received via a relay in Spamhaus SBL+XBL
>> tflags __RCVD_IN_SBL_XBL net
>>
>> I do have Mail::SpamAssassin::Plugin::URIDNSBL but maybe I need another plugin?
>>
>> What's missing?
>
>1) that's an RBL, not a URIDNSBL. You don't need anything but a reasonably
>recent version of Net::DNS
>
>2) That's a RBL parent test. You can pick it out by the fact that it's name
>begins with double underscore (__), which means it will run with no score.
>
>You'll never see it in a hit list for a message. Instead, you'll see it's
>children matching messages:
>
>header RCVD_IN_SBL eval:check_rbl_sub('sblxbl', '127.0.0.2')
>header RCVD_IN_XBL eval:check_rbl('sblxbl-notfirsthop',
>'sbl-xbl.spamhaus.org.', '127.0.0.[456]')
>
>The parent test exists to perform a single DNS query that the two children will
>later check against. This saves the overhead of doing a separate DNS query for
>each child.
Thanks for the breakdown! I had Net::DNS, but for kicks I
installed the CPAN version, no change. I've been happy with SURBL
but I'm not getting hits on SBL+XBL when I should be...
I've gone through configs and "skip_rbl_checks" is definatly not
set, nor do I see anything else that might turn it off. Do I have
to turn it on? Here is a sanitized debug log ('s/x/2' to get the
ip), I normally check in smtp so no user_prefs is okay.
2005-09-07 13:49:10.956908500 logmsg: connection from sta.duo [192.168.80.50] at port 33528
2005-09-07 13:49:10.960013500 debug: read_scoreonly_config: cannot open "/home/geo/.spamassassin/user_prefs": Permission denied
2005-09-07 13:49:10.960572500 debug: user has changed
2005-09-07 13:49:10.963970500 debug: Score set 1 chosen.
2005-09-07 13:49:10.975816500 logmsg: checking message <10...@eac71.neoplus.adsl.tpnet.pl> for geo:1002.
2005-09-07 13:49:10.986551500 debug: received-header: parsed as [ ip=83.x2.166.71 rdns=eac71.neoplus.adsl.tpnet.pl helo=eac71.neoplus.adsl.tpnet.pl by=sta.galis.org ident= envfrom= intl=0 id= auth= ]
2005-09-07 13:49:10.989047500 debug: received-header: relay 83.x2.166.71 trusted? no internal? no
2005-09-07 13:49:10.989055500 debug: metadata: X-Spam-Relays-Trusted:
2005-09-07 13:49:10.989059500 debug: metadata: X-Spam-Relays-Untrusted: [ ip=83.x2.166.71 rdns=eac71.neoplus.adsl.tpnet.pl helo=eac71.neoplus.adsl.tpnet.pl by=sta.galis.org ident= envfrom= intl=0 id= auth= ]
2005-09-07 13:49:10.990619500 debug: ---- MIME PARSER START ----
2005-09-07 13:49:10.991358500 debug: main message type: text/plain
2005-09-07 13:49:10.991804500 debug: parsing normal part
2005-09-07 13:49:10.992416500 debug: added part, type: text/plain
2005-09-07 13:49:10.992857500 debug: ---- MIME PARSER END ----
2005-09-07 13:49:10.993882500 debug: decoding: other encoding type (8bit), ignoring
2005-09-07 13:49:10.996434500 debug: URIDNSBL: domains to query:
2005-09-07 13:49:11.002893500 debug: is Net::DNS::Resolver available? yes
2005-09-07 13:49:11.005369500 debug: Net::DNS version: 0.53
2005-09-07 13:49:11.009737500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 58.
2005-09-07 13:49:11.026387500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 58.
2005-09-07 13:49:11.030191500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 58.
2005-09-07 13:49:11.044506500 debug: all '*From' addrs: affectionate@cyberpost.every1.net
2005-09-07 13:49:11.048007500 Argument "REPORT" isn't numeric in subroutine entry at /usr/share/perl/5.6.1/IO/Socket/INET.pm line 223, <GEN3> line 58.
2005-09-07 13:49:11.066033500 debug: Running tests for priority: 0
2005-09-07 13:49:11.066527500 debug: running header regexp tests; score so far=0
2005-09-07 13:49:11.151217500 debug: all '*To' addrs: geo-george@galis.org geo-geo-george@galis.org
2005-09-07 13:49:11.153705500 debug: forged-HELO: from=adsl.tpnet.pl helo=adsl.tpnet.pl by=galis.org
2005-09-07 13:49:11.162099500 debug: running body-text per-line regexp tests; score so far=-0.7
2005-09-07 13:49:11.466932500 debug: running uri tests; score so far=-0.7
2005-09-07 13:49:11.483724500 debug: Razor2 is not available
2005-09-07 13:49:11.495036500 debug: running raw-body-text per-line regexp tests; score so far=-0.7
2005-09-07 13:49:11.520367500 debug: running full-text regexp tests; score so far=-0.7
2005-09-07 13:49:11.522513500 debug: Razor2 is not available
2005-09-07 13:49:11.523012500 debug: DCCifd is not available: no r/w dccifd socket found.
2005-09-07 13:49:11.524709500 debug: DCC is not available: no executable dccproc found.
2005-09-07 13:49:11.525918500 debug: Pyzor is not available: pyzor not found
2005-09-07 13:49:11.526343500 debug: Running tests for priority: 500
2005-09-07 13:49:11.547875500 debug: RBL: success for 4 of 4 queries
2005-09-07 13:49:11.549006500 debug: running meta tests; score so far=-0.7
2005-09-07 13:49:11.560222500 debug: running header regexp tests; score so far=-0.7
2005-09-07 13:49:11.560761500 debug: running body-text per-line regexp tests; score so far=-0.7
2005-09-07 13:49:11.561137500 debug: running uri tests; score so far=-0.7
2005-09-07 13:49:11.570745500 debug: running raw-body-text per-line regexp tests; score so far=-0.7
2005-09-07 13:49:11.571165500 debug: running full-text regexp tests; score so far=-0.7
2005-09-07 13:49:11.571505500 debug: Running tests for priority: 1000
2005-09-07 13:49:11.571837500 debug: running meta tests; score so far=-0.7
2005-09-07 13:49:11.572461500 debug: running header regexp tests; score so far=-0.7
2005-09-07 13:49:11.573040500 debug: running body-text per-line regexp tests; score so far=-0.7
2005-09-07 13:49:11.573402500 debug: running uri tests; score so far=-0.7
2005-09-07 13:49:11.582738500 debug: running raw-body-text per-line regexp tests; score so far=-0.7
2005-09-07 13:49:11.583100500 debug: running full-text regexp tests; score so far=-0.7
2005-09-07 13:49:11.583878500 debug: is spam? score=-0.7 required=4
2005-09-07 13:49:11.584301500 debug: tests=GEO_TO
2005-09-07 13:49:11.584647500 debug: subtests=__CT,__CTE,__CT_TEXT_PLAIN,__HAS_MSGID,__HAS_SUBJECT,__HAS_X_MAILER,__HAS_X_PRIORITY,__MIME_VERSION,__MSGID_OK_DIGITS,__SANE_MSGID,__SARE_BODY_BLANKS_5_100,__SARE_BODY_BLNK_5_100,__SARE_HEAD_HDR_MIMEV,__SARE_HEAD_HDR_RCVD,__SARE_HEAD_MAIL_BAT1,__SARE_HEAD_MIME_VALID,__SARE_HTML_HAS_MSG,__SARE_HTML_HAS_TO,__SARE_SPEC_WATCH_BODY,__THEBAT_MUA,__THEBAT_MUA_V1
2005-09-07 13:49:11.594118500 logmsg: clean message (-0.7/4.0) for geo:1002 in 1.0 seconds, 1764 bytes.
2005-09-07 13:49:11.594660500 logmsg: result: . 0 - GEO_TO scantime=1.0,size=1764,mid=<10...@eac71.neoplus.adsl.tpnet.pl>,autolearn=disabled
// George
--
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org
Re: not using rbls?
Posted by Matt Kettler <mk...@evi-inc.com>.
George Georgalis wrote:
> it would seem the following rule is not being used...
>
> header __RCVD_IN_SBL_XBL eval:check_rbl('sblxbl', 'sbl-xbl.spamhaus.org.')
> describe __RCVD_IN_SBL_XBL Received via a relay in Spamhaus SBL+XBL
> tflags __RCVD_IN_SBL_XBL net
>
> I do have Mail::SpamAssassin::Plugin::URIDNSBL but maybe I need another plugin?
>
> What's missing?
1) that's an RBL, not a URIDNSBL. You don't need anything but a reasonably
recent version of Net::DNS
2) That's a RBL parent test. You can pick it out by the fact that it's name
begins with double underscore (__), which means it will run with no score.
You'll never see it in a hit list for a message. Instead, you'll see it's
children matching messages:
header RCVD_IN_SBL eval:check_rbl_sub('sblxbl', '127.0.0.2')
header RCVD_IN_XBL eval:check_rbl('sblxbl-notfirsthop',
'sbl-xbl.spamhaus.org.', '127.0.0.[456]')
The parent test exists to perform a single DNS query that the two children will
later check against. This saves the overhead of doing a separate DNS query for
each child.