[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: has anyone built a 'tripwire'-like ids detection system using svn's or other scm tools?

From: Kalin KOZHUHAROV <kalin_at_thinrope.net>
Date: 2006-01-18 01:32:11 CET

Urs Rau wrote:
> Here is another whacky 'off the wall' idea. And I promise that is the
> third and last of those sort of open ended questions, to this list.
>
> Has anyone thought off or tried building a set of scripts that would
> allow an admin to monitor and track remote unix servers (in my case
> mostly plesk webservers) and the files on them. I would be mainly
> interested in the actual OS files themselves, and maybe start watchign
> the html content of the websites on it as well, but that wouldn't be the
> primary use for this tool.
>
> The main aim would be to be able to easily and quickly see when a file
> has changed and what the exact change was (for config or text fles
> anyway). I would hope these script(s) would also allow me to monitor both
> binaries or executables and store such things like permissions as well as
> checksums as well as of course specified unix configuration text files.
>
> In some respects using scm tools like svn seems to be a much more
> appealing proposition to me, than using a tool like the free open sourced
> old version of 'tripwire'.
>
> The main advantage I can see is that with svn or any other scm tool used
> for this task, the admin would have the additonal help of having an exact
> record of what the file used to look like and what exactly has changed.
> And not jsut be told this has changed and then have to go figure why.
>
> And 'restores' or fixes fo problems could be done without having to go
> through the often cumbersome and long path of restoring from tape, or
> worse offsite tar file.

If you have a mirror of your installation on another box, it is not
cumbersome. Use `rsync -Hazv --dry-run to check host.domain:/path/ path/`

> PS: I am hanging around on as BigBear @ #svn at irc.freenode.net if you
> want to brainstorm these ideas or bounce comments around, or simply tell
> me off for asking these borderline to off-topic questions on the mailing
> list. ;-)

A very easy way to implement it is having a svn repository with your full
installation, then do a `svn export` on the live server. Then `rsync
--exclude=.svn/' to your working copy and use `svn stat` and `svn diff` for
that. Note that you will need at least four times the space (live server,
backup server x2 (for the pristine copy), svn server), but you will have 4
backups at any time in case you really screw it up. And don't do that all on
the same machine :-)

If you have the live server under svn and not use the backup server (just
_another_ server to host the repository) you can do it with 3 times the
space. But the 2 times are on the live server and you may not want that.

Kalin.

-- 
|[ ~~~~~~~~~~~~~~~~~~~~~~ ]|
+-> http://ThinRope.net/ <-+
|[ ______________________ ]|
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Jan 18 02:26:54 2006

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.