Bug#285554: [Pkg-nagios-devel] Bug#285554: acknowledged by developer (check_smtp update)

[Jeroen van Wolffelaar]

On Sun, Oct 09, 2005 at 04:40:37PM -0400, sean finney wrote:
> On Sun, Oct 09, 2005 at 06:15:18PM +0200, Jeroen van Wolffelaar wrote:
> > It sends a HELO, yes, but not an EHLO, which is still wrong when you want to
> > do TLS.
> > 
> > < 220 Personal mailserver of Jeroen van Wolffelaar <jeroen at wolffelaar.nl>, using Debian Exim
> > > HELO bla
> > < 250 mordor.wolffelaar.nl Hello 22pc220.sshunet.nl [145.97.220.22]
> > > STARTTLS
> > < 503 STARTTLS command used when not advertised
> > > ......................... at .......\c.zw...Vn.=.D^
> > < 421 mordor.wolffelaar.nl lost input connection
> > 
> > Note that nagios starts TLS even though it's being refused by the SMTP server.
> 
> okay, so then then that's two problems, as even though the starttls
> was refused to the first error, it blindly kept on going regardless of
> the 5xx error.

Yup, and the EHLO thingy being the primary one -- if all things go right, the
second one doesn't matter (but yeah, should still be fixed -- and result in a
nagios check failure).
 
> > It also doesn't send a FQDN as HELO/EHLO parameter, which is technically also
> > an SMTP protocol failure. the hostname -f FQDN would do, for example, and/or
> > making is configureable.
> 
> i don't know of a generalizable-but-portable way to get the 
> hostname+fqdn within C, but perhaps some ifdefs could be used to
> use gethostname/getdomainname if available for a default value,
> and provide a cmdline option to override that.  how does that sound?

I've no idea whether that'd work... on my system, I get '(none)' as
getdomainname() in any case.

what /bin/hostname does is basicly gethostbyname(hostname())->h_name. Without
any error checking, and prossibly dangerous C code:

#include <netdb.h>
main()
{
        char* hostname;
        struct hostent *hp;

        hostname = (char*)malloc(250);
        gethostname(hostname, 250);
        hp = gethostbyname(hostname);
        printf("%s\n", hp->h_name);
}
 
> > You are right though that it behaves a little bit better, it's just not quite
> > there yet. Something that, to be honest, should have been obvious reading the
> > buglog & peeking at the code(-changes), so I'm wondering why you closed the
> > bug, but oh well. In the even I suddenly get magically enough time left over,
> > I'll take a stab at the code myself... But honesty requires me to admit that's
> > quite unlikely to happen anywhere soon.
> 
> yeah, i suppose a more critical eye should have caught that, but given
> the large number of bugs i'm having to address w/nagios-plugins,
> i'm doing a bit of skimming here and there, as has been made evident.

Yeah, I noticed :).
 
> anyway, don't worry about spending time on this one, i'll hack a fix
> together.  however, if you could address the issues in the ntp-related
> bugs (verifying the old patch is still valid, and any additional
> patching/removal of code you'd recommend for the ntpdate stuff), i'd be
> very appreciative.

I'll reply to this while cc'ing the relevant bug instead of this one.

Thanks,
--Jeroen

-- 
Jeroen van Wolffelaar
Jeroen at wolffelaar.nl (also for Jabber & MSN; ICQ: 33944357)
http://Jeroen.A-Eskwadraat.nl