Not dismissing it. TCT is useful for forensic, for example a server
compromise. Yes I used it before but took a better deeper look at
mactime this time. I thought it could get the created timestamp for
files.

I think an accurate way to get the install date is by getting the
creation timestamp of the / partition. It is possible that some
journaling file systems has a record of the creation time in the
journal log. That is if the file system retains old information like
that because as far as I know most of them only record recent updates.

   Ed   <blog.eonsec.com>

On Dec 25, 2007 2:07 AM, Drexx Laggui [personal] <[EMAIL PROTECTED]> wrote:
> 25Dec2007 (UTC +8)
>
> On 12/24/07, Eduardo Tongson <[EMAIL PROTECTED]> wrote:
> > Took a peek. TCT mactime can not be used to determine the install
> > date. mactime uses lstat(2) which in turn relies on inode timestamps.
> > Inode timestamps only has one field each for modified, accessed,
> > changed so it can only record the last update.
>
> I suggest that you try mactime first before dismissing it immediately.
> It might help if you try not to look at specific and individual files,
> but to look at filesytem activity in its entirety. Mactime is *one* of
> the tools that can let you visualize (in your mind only) what happened
> on a filesystem, by listing all the file times in chronological order.
>
> And of course, mactime output shouldn't be the only source of
> information you should look at. You need to check out whatever
> available logs there are as well. You need to co-relate it with other
> data to get a definitive idea of what went on. And the timestamps from
> both the logs and mactime output should be cross-checked with the
> hardware BIOS date and time output. Individual file properties may
> mislead a system administrator, but shouldn't if the whole filesystem
> and contents were examined.
>
> Note: I don't suggest using mactime on a live fs. You'd have to dd the
> fs, mount the HDD image file with ro,loop,nodev,noexec,nosuid,,noatime
> options on a workstation, before running ils and fls, then actually
> running mactime. This is so that you minimize "dirtying" the
> filesystem being examined. Mactime output can be a big file, like a
> 40GB HDD can easily yield a ~95MB text file --and you'll be producing
> more than one.
>
> > I think the MAC letters are confusing. C can be mistaken for inode
> > creation timestamp. Like this Security book that mistakenly attributes
> > it. 
> > <http://books.google.com/books?id=xVuoTgSL1QoC&pg=PA620&dq=inode+timestamp&sig=Vg6TpwzxPVlqfVOX7pVUxL4Li-U>
> > Know your tools.
>
> For Windows NTFS users, ctime is "creation time". MAC times (modify /
> access / change; mtime / atime / ctime) for us in the Unix world, it's
> different. From stat(2): ctime is changed by writing or by setting
> inode information (i.e., owner, group, link count, mode, etc.). One
> perspective to take is to look at MAC times as reference to the file
> properties itself, not to the file content properties. To quote
> http://www.linux.com/feature/41179?theme=print
>
> "The atime attribute refers to the last time the file or directory was
> accessed. The mtime attribute, in contrast, changes when a file's
> contents are modified. The ctime attribute keeps track of when the
> contents or the meta-information about the file has changed: the
> owner, the group, the file permissions, and so on. The ctime attribute
> may also be used as an approximation of when a file was deleted."
>
> > On 12/24/07, Drexx Laggui [personal] <[EMAIL PROTECTED]> wrote:
> > >
>
> > > On 12/10/07, Federico Sevilla III <[EMAIL PROTECTED]> wrote:
> > > > On Mon, 2007-12-10 at 20:01 +0800, Drexx Laggui [personal] wrote:
> > > > >
> > > > > On 12/10/07, jan gestre <[EMAIL PROTECTED]> wrote:
> > > > > > I'm just after the install date.
> > > > >
> > > > > 'cat /proc/version' will give you the same output as "uname -a". The
> > > > > installation date is shown there.
> > > >
> > > > Caveat: /proc/version and `uname -a` provide you with the build date of
> > > > the kernel you are running. On systems where the kernel was upgraded
> > > > after the installation was done, this will not be an accurate measure of
> > > > the server's install date.
> > > >
> > > > Perhaps a more appropriate approach will be to try to find the change
> > > > date of the oldest system file (user files may have been extracted from
> > > > a tarball, inheriting the original timestamp... which while also
> > > > possible on system files is probably not as common). Again this isn't
> > > > fool proof, but it may be a bit more accurate when the kernel has been
> > > > modified.
> > > >
> > > > Federico Sevilla III
> > > > F S 3 Consulting Inc.
> > > > http://www.fs3.ph
> > >
> > > Thanks for the tip! "uname -a" or "cat /proc/version" is what is
> > > suggested on many first-responder guides on computer forensics. IIRC,
> > > it started with a CERT.org publication some years ago. Anyway, as
> > > noted by many already, there is not one "smoking gun" evidence that
> > > can give the answer right away, as a Linux system is a complex beast
> > > nowadays. The system analyst or admin must use a combination of tools,
> > > deduce the answer from all the data present, and arrive at a best
> > > possible conclusion.
> > >
> > > Another good tool to use is "mactime". Check out an article on how
> > > it's used here:
> > > http://www.linux.com/feature/41179
>
>
> Drexx Laggui  -- CISA, CISSP, CFE Associate, CCSI, CSA
> http://www.laggui.com  ( Singapore / Manila / California )
> Computer forensics; Penetration testing; QMS & ISMS developers; K-Transfer
> PGP fingerprint = 6E62 A089 E3EA 1B93 BFB4  8363 FFEC 3976 FF31 8A4E
> _________________________________________________
> Philippine Linux Users' Group (PLUG) Mailing List
> [email protected] (#PLUG @ irc.free.net.ph)
> Read the Guidelines: http://linux.org.ph/lists
> Searchable Archives: http://archives.free.net.ph
>
_________________________________________________
Philippine Linux Users' Group (PLUG) Mailing List
[email protected] (#PLUG @ irc.free.net.ph)
Read the Guidelines: http://linux.org.ph/lists
Searchable Archives: http://archives.free.net.ph

Reply via email to