Your message dated Sat, 16 Aug 2025 12:19:18 +0100
with message-id 
<CAJ3BuoR=E7OqJGq=sxvsgc8nhe7_6any6ltuvz0pxfrh4id...@mail.gmail.com>
and subject line Re: wontfix, but patches considered
has caused the Debian Bug report #215640,
regarding logcheck: documentation confusing/contradictory
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [email protected]
immediately.)


-- 
215640: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=215640
Debian Bug Tracking System
Contact [email protected] with problems
--- Begin Message ---
Package: logcheck
Version: 1.2.15
Severity: minor

Just in case you aren't already aware of this (I see the note about
documentation being updated as time permits):

I'm a bit puzzled after reading through the documentation on several points:
1) what is the order of loglevels from most to least output?
2) are ignore files from one level included in the ignore list for
other levels?  How?
3) What is the significance of the different types of security
warnings and how do they interact (193485 already covers some of this)?
/etc/logcheck has cracking, ignore, violations.  logcheck.conf refers
to ATTACKS, VIOLATIONS, and EVENTS.  How are all these related?
4) How should Debian system administrators modify the rules for their
own situation, to preserve maximum upgradeability?  (Perhaps the
information in README.Maintainer is relevant for this audience to, or
could be made relevant)?
5) How can the set of logfiles that is scanned be expanded or
controlled? (looks like logcheck.logfiles, but should be in the README
or man page)

The later items in the preceding list are just absent, but for the
first two there is contradictory material.  Here are some of the
relevant statements:

1)
README.Debian
------------------------
The reportlevel (that is, the degree of filtering applied to "routine"
system events) can be reduced from the default by changing the line:
        REPORTLEVEL="server"
to:
        REPORTLEVEL="workstation"
or the more secure:
        REPORTLEVEL="paranoid"
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This seems to say server has a fair amount of filtering, workstation
less, and paranoid the least amount of filtering.  So in terms of
output:
server < workstation < paranoid 

But README.logcheck says "paranoid" is "high verbosity," server
intermediate, and workstation the least amount of output:
workstation< server < paranoid

A priori the order isn't clear to me: servers are more likely to be
attacked, and have more routes open for attack, so you might want more
logs with them.  But they also have more routine activity that you
wouldn't want to know about, so maybe their logs should have more
filtered out.

2) README.logcheck says (in REPORTLEVEL section) that workstation
applies all filters in workstation, server, and paranoid.  It says
server applies all in server and paranoid, and paranoid only does
paranoid.

That makes it sound as if the program automatically merges the
different directories, as appropriate.

README.Maintainer says "As the higher level ignore.d directories
include the lower levels (i.e. server = server + paranoid) you should
try to split your rulefile between the different ignore.d
directories."  

That makes it sound as if the directories are somehow combined on the
disk by the directory structure or symlinks.  It's unclear whether
"higher level" refers to higher level in directory structure, higher
level of logging, or higher level of filtering.

It adds "Packages should not install symlinks between
the ignore.d directories but install separate files into each level;
note that it is no longer necessary anyway for rules to be repeated in
each ignore.d.* directory."

That says it's not symlinks, and sort of sounds as if the 3 levels
cumulate because the same files appear in each of the directories.  In
other words, they cumulate because of the way the files happen to be
defined, and they might not cumulate if they were redefined.  But then
it says they don't need to be repeated, so it's uncertain what it
means.

Obviously these discussions also bear on which set of filters is most
restrictive (question 1).

Thanks for your work on the package.

-- System Information:
Debian Release: testing/unstable
Architecture: i386
Kernel: Linux iron 2.4.20 #1 SMP Thu Jun 26 15:57:25 PDT 2003 i686
Locale: LANG=C, LC_CTYPE=C

Versions of packages logcheck depends on:
ii  cron             3.0pl1-81               management of regular background p
ii  debconf          1.3.14                  Debian configuration management sy
ii  debianutils      2.5.5                   Miscellaneous utilities specific t
ii  exim4-daemon-lig 4.22-5                  Lightweight version of the Exim (v
ii  lockfile-progs   0.1.9                   Programs for locking and unlocking
ii  logcheck-databas 1.2.15                  A database of system log rules for
ii  logtail          1.2.15                  Returns parts of logfiles that hav
ii  mailx            1:8.1.2-0.20030521cvs-1 A simple mail user agent
ii  sysklogd [system 1.4.1-10                System Logging Daemon

-- debconf information:
  logcheck/changes: 
* logcheck/install-note: 



--- End Message ---
--- Begin Message ---
On Sun, 10 Nov 2024 18:51:15 +0000 Richard Lewis
<[email protected]> wrote:
> On Sat, 21 Oct 2006 18:26:44 +0200 martin f krafft <[email protected]> wrote:
> > tags 215640 help wontfix
> > thanks
> >
> > I won't fix this but will gladly consider patches.
>
> I think this should just be closed, the manpage has been amended in
> the last 20 years
>
>

closing on this basis. If there are other improvements to the
documentation, please open a new bug with the details and i will
gladly help

--- End Message ---

Reply via email to