in reply to WORM: Warning to all LINUX/PERL monks

It is a bad worm, yes, and it's taking advantage of a problem that many linux users (but doubtfully the ones on PM have) and that's the lack of updating critical software. The worm's based off a bug found in January for the so-called Raman worm which had to be used in conjunction with wuftpd and one other client, and affect all versions of BIND except the most recent, 8.2.3 (non-beta). Nearly all major *nux vendors had patches out the same day, but quotes estimate that maybe only 50 to 75% of those running *nix took this necessary step.

This worm is using the exact same exploit, delievering a much deadlier protocol since it basically sticks a root kit on the affected box. And it's exploiting the fact that some *nix installs NEVER get patched for security holes (one article I read says that 20% of the internet is affected because of the number of *nix-based servers that have old BIND versions. But this is why I don't think most PM-ers will be affected, because we are computer professionals and know the value of security patches :D.

(Of course, the other trick besides patching is to stick the bind process behind a different user besides root, which means that a would-be attacker would not be able to root-kit your machine.)


Dr. Michael K. Neylon - mneylon-pm@masemware.com || "You've left the lens cap of your mind on again, Pinky" - The Brain
  • Comment on Re: WORM: Warning to all LINUX/PERL monks

Replies are listed 'Best First'.
Re: Re: WORM: Warning to all LINUX/PERL monks
by Clownburner (Monk) on Mar 24, 2001 at 22:55 UTC
    Although it doesn't apply to this worm really, even if BIND is not run as root, it usually starts that way, so if you could exploit this hole, you could plant a "bomb" that would effectively root the box the next time BIND was run as root (such as a restart), which could then be provoked with any number of DoS attacks...

    But not running BIND as root is certainly the first step. What we need is a clean-slate, audited-code DNS implementation that is configuration-file compatible with BIND -- That would certainly make migration easier for the rest of the world, and maybe we'd get away from these BIND-related nasties for good.


    Signature void where prohibited by law.