On Thu, December 28, 2006 4:25 pm, Erik de Castro Lopo wrote:

>> perhaps, but, it certainly would have prevented two infiltration I had
>> in the last few month
>
> I *REALLY* honestly don't think so. Once somebody is in they will
> use a Perl script. If Perl isn't installed they will do something else,
> like uploading a precomiled binary.
>
> The idea is to prevent people getting in to begin with. Once they
> are in its way too late.

Erik, thanks

you obviously have a great deal more experience than I do, and, I agree
the idea is to prevent in the 1st place, however,

going by a rather limited experience of two semi-successful infiltrations,
they call a code from a remote server like 'sg-webDOTorg/cmd2DOTtxt'
(change DOT to .), trying 5 downloaders until it executes,

I'm guessing if the download/execution fails, the scripts will
search/target other servers

sure, what I'm suggesting will not stop a serious attempt to exploit a
hole, but, it should deflect such a script

as it was, when I realized the server was infiltrated, the 'solution' was:
remove downloaders, remove perl, reboot server, problem removed;
next day the problem was located, 'faulty' CMSs were deleted, and, Perl
re-instated

so, until the CMSs were removed, someone could've run different exploits,
but it didn't happen.

I feel this is like moving ssh to a non-stand port, a small measure to
reduce exposure.

lastly, now that '/tmp' is mounted as
/tmp type ext3 (rw,noexec,nosuid,nodev,noatime,nodiratime)
that should hopefully prevent execution of such expolits

thanks for all the comments

-- 
Voytek

-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Reply via email to