'failed setting up socket' error with BIND9 on Solaris 2.5

Gregory Hicks ghicks at cadence.com
Fri Dec 2 13:56:49 UTC 2005


> Date: Fri, 2 Dec 2005 14:35:55 +0100
> From: Udo Zumdick <uz at nic.dtag.de>
> To: BIND-Users Mailinglist <bind-users at isc.org>
> Subject: Re: 'failed setting up socket' error with BIND9 on Solaris 2.5
> 
> 
> Dear List,
> 
> apart from the xfr problems I had with BIND under solaris 2.5, which are solved
> with the help from Mark, the BIND 9 is still not working properly. This error
> was related to some kernelbugs in solaris 2.5

Upgrade to a more modern OS.  Solaris 2.5 is WELL past its "expire by" date.

(Solaris 2.5 was EOLd sometime ago...)

> 
> Beside this: Are there any solaris patches to solve this without modifying
> bind source ? I couldn' find a patch that helps.
> 
> After modifying the source as Mark suggested the xfr works good. But although
> there are no recursion restrictions in the configuration the BIND does not
> work as a resolver should do. Obviously there are some (more many) recursive
> queries that will not be resolved by this bind installation. I could not
> figure out which queries will be answered and which not and why.
> It only works with an external forwarder.
> 
> There are no network issues or something like this. Also the root.cache is
> up to date and works in other bind installations.
> 
> I tried out bind 9.2.5 and 9.3.1
> 
> I've never had such problems with any version of bind.  ?-)
> 
> Is it possible, that a BIND9 does not work with a Solaris 2.5 ? I can't
> imagine this...
> 
> 
> Thanks a lot for help !
> 
> -Udo 
> 
> 
> On Tue, 22 Nov 2005 09:09:36 +1100
> Mark Andrews <Mark_Andrews at isc.org> wrote
> > 
> > > Hi Mark,
> > > 
> > > > 	Looks like you are dealing with a kernel bug in SunOS 5.5.1.
> > > > 	My bet is that connect() is complaining that bind() has already
> > > > 	been called on the socket.  
> > > 
> > > Yep, you're right ! :-)
> > > 
> > > > This is done to get the correct
> > > > 	source address for the packets.
> > > > 	You should be able to remove the isc_socket_bind() call but
> > > > 	transfer-source will nolonger be effective for the TCP connection.
> > > [....]
> > > 
> > > 
> > > I compiled the bind again with your suggestions and that seems to work for me
> > > now. I am really happy :-)
> > > The transfer-source option is  not really needed in my case.
> > > 
> > > Thanks a lot for your help ! 
> > > 
> > > -Udo 
> > 
> > 	No worries.
> > 
> > 	Note there will be a similar issues with the isc_socket_bind()
> > 	calls in lib/dns/resolver.c and lib/dns/request.c.
> > 
> > 	Mark
> > --
> > Mark Andrews, ISC
> > 1 Seymour St., Dundas Valley, NSW 2117, Australia
> > PHONE: +61 2 9871 4742                 INTERNET: Mark_Andrews at isc.org
> 
> 

---------------------------------------------------------------------
Gregory Hicks                           | Principal Systems Engineer
Cadence Design Systems                  | Direct:   408.576.3609
555 River Oaks Pkwy M/S 6B1             | Fax:      408.894.3479
San Jose, CA 95134                      | Internet: ghicks at cadence.com

I am perfectly capable of learning from my mistakes.  I will surely
learn a great deal today.

"A democracy is a sheep and two wolves deciding on what to have for
lunch.  Freedom is a well armed sheep contesting the results of the
decision." - Benjamin Franklin

"The best we can hope for concerning the people at large is that they
be properly armed." --Alexander Hamilton



More information about the bind-users mailing list