ot, dig, rndc, nslookup segfaulting on debian

[Gregory Nowak]

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

Hi folks.

I'm running into a very weird problem, and am at a loss as to what to
do about it,besides what I've already done.

I am running a user-mode-linux machine
<http://user-mode-linux.sourceforge.net/>. The kernel is linux
2.6.12.5, and the distro is debian sarge stable. I'm running stable,
because this machine will act as a server once I've got everything
installed and configured, and I don't want the dependency problems that
can result when running testing from my previous experiences. 

Anyway, I've installed bind9, which installs bind 9.2.4, and all the dependencies (libisc7,
libisccc0, libisccc0, and libdns16) as well as bind9-host, dnsutils,
and bind9-doc. As the subject says, running dig, rndc, and nslookup
causes them all to produce a segmentation fault, though named itself
does start without problems. If I do something like 

dig -h

or

rndc

I get the usage message, and the programs exit normally. However, if I
do something like

dig yahoo.com.

rndc status

the only thing I get after hitting enter is "Segmentation fault",
followed by the prompt.

If I do

nslookup -h

I get

Invalid option: h
Segmentation fault

Just doing

nslookup

produces

Segmentation fault

So, I made another copy of my debian file system, so I could play
around, and upgraded that to testing, which still gave me the same
results as with stable.

So, then I went back to my file system with debian stable on it, and
used apt-build to build bind9 and the libs from source, but those
packages give me the same thing. For the heck of it, I then installed
the bind package, which installs bind 8.4.6 and dependencies, but got
the same thing with that.

Here's the weird part, this isn't a kernel issue, or a debian sarge
issue. I also have a slackware 10.1 file system here, and I run that
using user-mode-linux (uml). The version of bind that comes with slack
10.1 is bind 9.3.0, and dig, rndc, and nslookup run just fine on that
file system using the same exact uml kernel that I used with the
debian file systems.

I have an extra physical machine that I put debian sarge on a few
weeks ago for testing purposes. So, I fired up that box, did apt-get
install bind9, which gave me all the packages I needed. Guess what,
dig, rndc, and nslookup run on the physical machine, running debian
sarge, just fine.

I tried to run dig through gdb, but was told that no debugging symbols
were found, which didn't surprise me. Given that 3 different programs
are doing this, from 2 different packages (bind9 and dnsutils), I
suspect there's a library issue somewhere, but really don't know at
this point what I should do next. I guess I can extract the binaries
and libs from the slackware tgz package, but I don't know how I'd
turn them into a deb package, and I don't want to install the stuff
manually by just dumping it into /usr/local. I also would like to keep
compatible with debian by storing the configs into /etc/bind, instead
of into /etc, and /var/named the way slackware does. So, if anyone is able to please,
please help, that would be very much appreciated, and thanks in
advance.

Greg


- -- 
web site: http://www.romuald.net.eu.org
gpg public key: http://www.romuald.net.eu.org/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)

- --
Free domains: http://www.eu.org/ or mail dns-manager at EU.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFDE4Tq7s9z/XlyUyARAtoPAJ9CQoBsChRSWtGK9MLHQx5+bFZSzQCeP+PI
0Q9VnFUZWT6IfX60PPX/C1s=
=6dA/
-----END PGP SIGNATURE-----