All of lore.kernel.org
 help / color / mirror / Atom feed
From: Steve Grubb <sgrubb@redhat.com>
To: John Dennis <jdennis@redhat.com>
Cc: linux-audit@redhat.com
Subject: Re: Configuration file monitoring - reporting content changes
Date: Tue, 21 Jul 2015 13:59:01 -0400	[thread overview]
Message-ID: <3307081.aEimE15yPr@x2> (raw)
In-Reply-To: <55AE5967.70609@redhat.com>

On Tuesday, July 21, 2015 10:38:31 AM John Dennis wrote:
> On 07/20/2015 07:08 PM, Steve Grubb wrote:
> > On Monday, July 20, 2015 09:53:47 PM Burn Alting wrote:
> >> I am interested in any Linux based capability that will monitor
> >> identified files and report on actual changes to the monitored file.
> > 
> > I know of nothing that does this. But as long as the list of files is
> > limited, it doesn't sound like a hard program to write.
> > 
> > Any one else with an opinion?
> 
> Yes :-) I'm not so sure it's an easy program to write and be robust in a
> variety of scenarios. I know because I wrote such a program once. The
> basic problem is most people think in terms of monitoring a file by name
> (e.g. it's pathname). But inotify operates on inodes, not filenames. If
> that file is subject to any variety of log rotation strategies or
> modifications by a configuration manager whereby the file is renamed or
> moved to a different directory then any program using inotify to monitor
> the file needs to become reasonably sophisticated and be able to track
> those changes. It is entirely possible for two processes to have opened
> the same file by name but have them be 2 different files (e.g. after
> opening the file path is modified but the process still has the original
> inode open, now a 2nd process opens the same filename but gets a
> different inode). Conflating inodes with filenames can lead to
> unexpected results and if the purpose is some sort of security
> monitoring it will be important these issues are accounted for.

I recently was doing some experimenting with the fanotify API. In my mind, I 
think its likely to be better. But it has limitations such as mmap'ed file may 
not generate a modify event. So, if I were going to do it, I'd start there. 
But you do raise a whole lot of good points. My guess is this would watch 
config files which logrotate wouldn't apply. But yes, editors do open a temp 
copy and then do a rename. In my experimenting, I didn't bother to see how 
fanotify handle renames. (You would think its a modify event.)

-Steve


> Some of this is discussed in these documents which accompany the lwatch
> (Log Watch) program I wrote:
> 
> https://jdennis.fedorapeople.org/lwatch/html/InotifyOverview.html
> https://jdennis.fedorapeople.org/lwatch/html/LogWatchOverview.html

  reply	other threads:[~2015-07-21 17:59 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-07-20 11:53 Configuration file monitoring - reporting content changes Burn Alting
     [not found] ` <201507202109.GAJ05274.tSFOJFFVQMLOHO@I-love.SAKURA.ne.jp>
2015-07-20 13:03   ` Burn Alting
2015-07-20 17:53 ` Smith, Gary R
2015-07-20 22:12   ` Burn Alting
2015-07-20 23:08 ` Steve Grubb
2015-07-21 13:47   ` EXT :Re: " Boyce, Kevin P (AS)
2015-07-21 21:48     ` Burn Alting
2015-07-21 14:38   ` John Dennis
2015-07-21 17:59     ` Steve Grubb [this message]
2015-07-21 21:54       ` Burn Alting

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=3307081.aEimE15yPr@x2 \
    --to=sgrubb@redhat.com \
    --cc=jdennis@redhat.com \
    --cc=linux-audit@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.