Penetration Testing mailing list archives

Re: [PEN-TEST] Load Balancing Servers


From: Miller Scott Contr 30CS/FTI <Scott.Miller () VANDENBERG AF MIL>
Date: Wed, 27 Sep 2000 08:50:37 -0700

I'd have to disagree with you there.  Just because the servers are in a load
balancing configuration and they appear to have the same web content doesn't
mean they're identical in their vulnerabilities.  If reasonable
configuration control measures are taken, they should be, but we all know
how often that's not the case.  Remember, our jobs would be a lot harder
were it not for lazy or overworked sysadmins.  Take DNS zone transfers, for
example - while the data is the same on each authoritative server, I've
found that in some cases one server will be misconfigured to allow
unauthorized zone transfers while the others will be properly secured.  It
pays to check every system you can access.

What about identifying the server you're connected to?  From working with
IIS webservers, I would think it should be possible to get some useful
information from the HTTP headers returned.  Has anyone tried this in
practice?

Scott

-----Original Message-----
From: Jens Knoell [mailto:jens () ING TWINWAVE NET]
Sent: Wednesday, September 27, 2000 8:02 AM
To: PEN-TEST () SECURITYFOCUS COM
Subject: Re: Load Balancing Servers

That's not entirely true... the problem is more that you cannot select which
server you attack. Assuming that load balancing servers are all mirrors of
each other, it shouldn't really matter what you test. Find a
vulnerability/weakness on one server, and you can more or less assume you
got the same problem on the mirrors too.

Jens


Current thread: