[lug] resolver problems

D. Stimits stimits at idcomm.com
Fri Mar 2 11:51:09 MST 2001

John Hernandez wrote:
> Whenever I use dig, host, or nslookup and specify a nameserver other than the default, I get a Segmentation fault.  But it only happens if I use FQDN rather than IP address for the server specification.  I tried strace, but it didn't really help me.  It produces a core file, but I'm not experienced in corefile analysis either; completely clueless what to do after 'gdb dig core'.

Assuming it was compiled with "-g" for debugging, you'll find more of
what you need. But without that, you might still find something useful.
A core file should set memory up the way it was at the time of death, or
you could simply run "r" and get the same results. If you need arguments
before running, you can use "set args whatever" first. But the basic
info would be a stack frame (via back trace "bt"), and possibly a list
of lines surrounding the current line via "l".

On the other hand, you might get more information by determining what
library versions it has available, and which ones actually exist.
Debugging info will find out where a problem is, but it won't tell you
if it is due to a bug or an incompatible version. cd to the app
directory, and use ldd on it: "ldd whatever". Then compare it to the
list of libraries your system has, as root: "ldconfig -p | sort | less".

> My system:
> Linux 2.4.1
> RH 7.0.99 (fisher)
> i686
> $ dig @name.ufl.edu www.ufl.edu
> Segmentation fault
> $ host name.ufl.edu
> name.ufl.edu. has address
> $ dig @ www.ufl.edu
> ; <<>> DiG 9.1.0 <<>> @ www.ufl.edu
> ;; global options:  printcmd
> ;; Got answer:
> I found something about this in Redhat's bugzilla system, and the engineer's advice was "upgrade your glibc".  Well, I can't find a more recent glibc, and there are no updates in the fisher tree.
> $ rpm -q glibc
> glibc-2.2.1-3
> Any ideas?
> -John
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug

More information about the LUG mailing list