pull-request.
This would help developers most.
--
Schönen Gruß
Hartmut Goebel
Dipl.-Informatiker (univ), CISSP, CSSLP
Goebel Consult
http://www.goebel-consult.de
Monatliche Kolumne:
http://www.cissp-gefluester.de/2012-02-bring-your-own-life-glosse
Blog: http://www.goebel-consult.de/blog/2005
e applied and should go into 1.2.1.
According to RFC 1034, sec.3.5, the hyphen character ('-') is a valid
character in hostnames. If shinken doe not support this, this is a major
annoyance IMHO.
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
have
learned a lot of it and am using this model for my projects. I recommend
using this model (or a slightly adopted version).
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compilers.com | compilers which you thought ar
ore than
it seams to have use.
What do you think about this proposal:
- for brok-messages: no human format at all
- for console-outpur: always human format
- for local-log: always human format
An I'll change the human format to the standard f
to wrote a "dump all
> logs" module in the broker side) and if the module is smart enough to
> split them, it just got to look at the brok :)
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe.
, too, so each state-change would be send twice.
IMHO we do not need to log state-messages at the scheduler only with
level "DEBUG" which is not send to the broker. State-messages are not
interesting at the scheduler side except for debugging.
What do you think?
--
Regards
Ha
5) Do we need 'human format' for messages send as broks, too?
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compilers.com | compilers wh
Am 22.08.2012 15:44, schrieb Francois Mikus:
> This may not be feasible because we already have a large test library in
> /test.
Doctests could provide an additional way for testing. If you find
places, where doctests are useful, go ahead!
--
Regards
Hartmut Goebel
| Hartmut
mething else before posting...
So this option should be available on platforms supporting it.
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compilers.com | compilers which you thought ar
does not show up errors, I'll continue with the new
logging module.
So please be patient just a few more days :-)
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compilers.com | compilers which you thou
g only
> interesting stuff. Debug has way too much information that
> can
There may (I don't know yet) be better ways to achieve this.
Before implementing, please what some days. I'm already working on a new
logging module, but need to implement performance-tests
w to write a good commit message".
I would really appreciate to see more meaningful commit messages in the
future.
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compilers.com | compi
crete default
message.
The switch could be build into shinken_test.py quite easy, so developers
working with 2.7 do not need to install the backport.
What do you think? What is requires for this, esp. to keep the
Jenkins-tests running?
--
Regards
Hartmut Goebel
| Hartmut Goebel |
Am 27.07.2012 21:23, schrieb Hartmut Goebel:
> "Convince me" is not such a good argument if you want to attract
> developers :-\
I mean: "Convince me why I should follow the standard" is not such a
good argument if you want to attract developers :-\
"Convince me
e the Python way to do it, see PEP 8 and PEP 257.
"Convince me" is not such a good argument if you want to attract
developers :-\
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compilers.com | compilers
#99.
Unfortunately Shinken is lacking here.
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compilers.com | compilers which you thought are impossible |
--
L
eriods, to name
just a few.
As you can see, my cleanup is not about maintaining configs, but about
keeping them tidy.
> So
> keeping all N distinct configuration files is even a good thing for tests,
> especially regression check.
This would require the test-case to check the regression. :
crash :p ).
It would be great if you could hand over some examples or test-cases as
test_logging.py is very, very basic ATM and does not really test
anything (but just output).
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compi
ease.
> No Shinken in OMD for CentOS/RedHat 5.x which is a very bad idea (it's
> already a castrated package because it includes no Shinken UI for these
> versions)
Yet another reason why OMD should include a recent Python version :-)
--
Regards
Hartmut Goebel
| Hartmut
kets).
Is this intended? Or just caused by nobody cleaning this up until now?
Is it okay if I'm going to remove the second timestamp and change this into
2012-07-15 20:12:48,397 Info: Some log-message
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com
Hi,
I'm wondering if there is any reason for importing `objects` in
shinken.__init__. Thus *every* script or test-case importing shinken
gets all the objects loaded, even if they are not required by the script
or test-case at all.
Can we remove this?
--
Regards
Hartmut Goebel
| Hartmut G
which are usefull for not reinventing the wheel.
So if there is not reason for using a own implementation I'll change to
a standard-implementation. I would do this in a separate branch for you
to test and comment on it.
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-com
Hi,
the WebUI is based on "bottle". As I just discovered, bottle requires
Python 2.5 or newer.
AFAIK The official minimum requirement for Shinken is still Python 2.4.
I propose pushing the official minimum requirement for Shinken to at
least 2.5, too.
Any opinions?
--
Regards
Hart
--
Schönen Gruß
Hartmut Goebel
Dipl.-Informatiker (univ), CISSP, CSSLP
Goebel Consult
http://www.goebel-consult.de
Monatliche Kolumne:
http://www.cissp-gefluester.de/2012-02-bring-your-own-life-glosse
Blog: http://www.goebel-consult.de/blog/20050315
Goebel Consult ist Mitglied bei http:
The "multiple spaces after ':'" mostly originate from formatting
properties into "tables". THis results in more readable code, but is not
strictly PEP 8.
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.cra
test some
> latin-9 caracter in comments for example)...
Okay, So I changed the code to contain the same characters, but in
octal-notation (\351) instead of a single byte.
Output is still the same as the one you showed.
--
Schönen Gruß
Hartmut Goebel
Dipl.-Informatiker (univ), CISSP, CSS
octal 351 == int 233 == hex e9). In
Latin-1, e9 is small letter e with accute.
What is the intention of this line? What should be tested?
--
Schönen Gruß
Hartmut Goebel
Dipl.-Informatiker (univ), CISSP, CSSLP
Goebel Consult
http://www.goebel-consult.de
Monatliche Kolumne:
http://www.cissp-geflues
would
change the file and strip leading/trailing white-spaces which would make
the file format more robust.
--
Regards
Hartmut Goebel
| Hartmut Goebel | h.goe...@crazy-compilers.com |
| www.crazy-compilers.com | compilers which you thought are impos
_install.
Thanks.
--
Schönen Gruß
Hartmut Goebel
Dipl.-Informatiker (univ), CISSP, CSSLP
Goebel Consult
http://www.goebel-consult.de
Monatliche Kolumne:
http://www.cissp-gefluester.de/2011-09-kommerz-uber-recht-fdp-die-gefaellt-mir-partei
Blog: http://www.goebel-consult.de/blog/20050620
G
on it)?.
--
Schönen Gruß
Hartmut Goebel
Dipl.-Informatiker (univ), CISSP, CSSLP
Goebel Consult
http://www.goebel-consult.de
Monatliche Kolumne:
http://www.cissp-gefluester.de/2011-10-aus-der-schublade-in-die-koepfe
Blog: http://www.goebel-consult.de/blog/20060215
Goebel Consult ist Mitglied
/launch_all.sh
@nap: we should fix the documentation. Installation should be done using
setup.py instead of copying.
I also tried to start shinken as root :
You must not run shinken as root for securiy reasons.
(I'm thinking about building a save-guard into shinken.)
--
Schönen Gruß - Regards
Ha
d I miss something?
> If we comment it quite good, it's not a howful hack after all, and
> it's quite even clear what we do (just set in the path so we can just
> import shinken/ after all), we should add such a comment for imp too
> after all.
ACK for comments.
--
Schönen
ot;, ".."]))
This doe not change sys.path, but tries to import shinken from the
current directory and the parent directory. This is what we want at
last, isn't it?
What do you think about this?
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Go
ul is
absolutely okay for me.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitglied bei http
hold a list -- this contradicts teh
semantics expressed by the name "StringProp".
So again my question: What type of property is this?
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebu
7;s all. So we can add a new property class,
> rawobject or something like it, for such object we do not want to
> "Pythonize".
IMHO this should tell something about the type of the property, So
"RawProperty" is not a good name.
What are these values used for? Why th
e explain this to me?
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitglied bei http://www.7
Hi,
I just checked in some fixes for
Changing `xx ==/!= None` into `xx is (not) None` as suggested by PEP 8.
This has been done using another fixer based on lib2to3. So there should
be semantic change within this.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP
IC. But in this case a verbose prfix like "python:" or "py:" would be
much better and esier to understand.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel
user and *one* group. This is okay
for me.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult m
derstand, speaks for itself, handles
alternate separators and is shorter, too.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cis
job
to set these ownerships? Isn't this better left over to the packagers?
Awaiting your comments.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatlich
ill in the script:
- some paths are hard-codes (e.g. for all "data_files"
- user and group is hard-coded
My plan is to implement some more "sub-commands" (distutils terminology)
to solve these issues. Hopefully I have time this weekend.
--
Schönen Gruß - Regards
Hartmut Go
quot;module_type"
there already is a obvious way to do this (which is more flexible, too)
So this feature does not add any functionality but only add future
longstanding pollution.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spez
Am 10.02.2011 13:31, schrieb Grégory Starck:
> since the commit with distutils the "setup install" (&
> test_end_to_end) are broken on my side :s
Thanks for your report I'll look into it.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CIS
Am 10.02.2011 19:57, schrieb Hartmut Goebel:
> Am 10.02.2011 13:36, schrieb nap:
>> Yes, it's a regression bug. We must get back the setup_parameters.cfg
>> because packagers need it. I think with it quite a lot of the above
>> bugs will disapears.
> IMHO we should fix
just
a small change.
setup_parameters.cfg has the big, big disadvantage of being non
standard. Why use an additional config file if distutils already have
one and read it automatically?
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sich
server restart.
>
> So it's a waste of time to operate in production and it takes longer
> to package.
>
> Is it possible to return to the startup script are complying with the LSB?
Can you please explain, what's the exact problem, the "bad idea". Then
we can fix it
sion_info
(2, 6, 5, 'final', 0)
>>> sys.version_info < (2,7)
True
>>> sys.version_info > (2,6)
True
# ough!
>>> sys.version_info < (2,6)
False
>>> sys.version_info <= (2,6)
False
>>> sys.version_info >= (2,)
True
>>> s
r rc versions and look if they keep the + or not. From
> now, just comment this pythjon_version checking, 2.7 is good.
IMHO the best solution is to drop this useless platform.python_version
thing and simply use sys.version_info.
I'll prepare commit a patch.
--
Schönen Gruß - Regards
Hartm
t later.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitglied bei http://www.7-it.de
smi
Am 07.02.2011 18:48, schrieb Gerhard Lausser:
> Dead since 2008 after a few weeks of activity. The author probably found
> out that it's not so easy to implement the native Java RMI in Python :-)
So JMX is using RMI. Then this not a useful approach.
--
Schönen Gruß - Regards
Hartmut
Shinken.
Oh my God! PyPerl is a zombie! Stay away from it!
IMHO it far less trouble rewriting the plugins in Python (using the new
plugin base classes), than trying to make perl load into the Python process.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goe
ller in Java, just to talk JMX.
There is <http://code.google.com/p/pymx/>
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.c
sntallations.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitglied bei http://www.7-it.de
Am 04.02.2011 13:59, schrieb nap:
> We can add the logging with simple rotation options, maybe the same
> that the simple_log module, so users won't be lost.
Yes, and we can reinvent the weel. SCNR
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goe
te-wmi #
So a untested idea. And nothing to implement soon.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goe
ng yet-another-mechanism-for-log-config.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitgli
Am 04.02.2011 11:37, schrieb nap:
> use_local_log=0
> local_log=%(workdir)s/pollerd.log
Ough, this hurts!
Why are you not using the default configurations options from the
standard logging module?
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel C
you are on the save side
(well, German proverb: auf der sicheren Seite)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-geflues
a resp. bug regarding multiprocessing on windows:
<http://bugs.python.org/issue5162>. In multiprocessing AFAIK no new
interpreter is started, which makes me believe, such a bug is possible ;-)
BTW: No offence meant :-)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.),
Am 03.02.2011 11:46, schrieb nap:
>
>
> On Thu, Feb 3, 2011 at 11:40 AM, Hartmut Goebel
> mailto:h.goe...@goebel-consult.de>> wrote:
>
> Am 03.02.2011 11:32, schrieb nap:
> > For the sys.path, it's a bug I've got in centos 5, python2.4, direct
>
What? Are you kidding?
subprocess starts a new python interpreter which runs a new main script
which should have a __file__ attribute then. I do not see, why this
should not happen? Did you file a bugreport at python.org?
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CS
is there.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitglied bei http://www.7-it
ot; directory.
>
> Or am I missing something.. :-?
I can't see, why this should be needed, too.
Further: I can not imagine why the __main__ module could not have a
__file__ attribute. (At least if __main__ is a Python script.)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (
s) and a Result() object for returning the data.
The major benefit for such a result object is that it can be processed
quite easily, ads the dummy example for check_multi shows: Results are
more or less simply added, no need to parse a string.
Comments are welcome.
--
Schönen Gruß - Regards
H
ner, poller, etc.
If someone wants to deploy a poller to a system then, he (ideally) only
put there shinken/common and shinken/poller. If we may introduce
namespace packages somewhen, this could even become something like
easy_install shinken.poller shinken.plugins.oracle shinken.plugins.hpux
That&
t, since I did not understand the code
good enough ;-)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goeb
t an idea yet. I'm not going to
implementent this now.)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel
st()-Method is
meant for.
>> * The performance data will be a dict. Keys are strings, values are
>> either integer, float or string (percent), or a list/tuple thereof.
> What about min/max/thresholds?
I thought, simple lists of values are enough. Am I missing something?
--
Schön
single "namespace".
> (by the way, argparse is widely package for major distributions)
I don't understand this sentence.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goeb
code of the check needs to be changed, the
Open questions:
* Should we use optparse or argparse? optparse is standard since Python
2.3 (AFAIR). argparse is standard staring with Python 2.7, but is
available for 2.4, too (AFAIK). argparse has a much cleaner interface.
--
Schönen Gruß
ules with a good "Class API".
You are right. In addition a class enables some other possibilities
without change to each "check". E.g. implementing threaded execution on
choice of the application.
Maybe I'll have time this weekend so I can implement an example as a
base
#... parse args
#... put a lot of magic here
#... and run check
if __name__ == '__main__':
testrun()
You may want to have a look at
<http://svn.origo.ethz.ch/wsvn/pdfposter/trunk/pdftools/pdfposter/cmd.py> and
<http://svn.origo.ethz.ch/wsvn/pdfposter/trunk/pdftools/pd
Am 19.01.2011 16:57, schrieb Gerhard Lausser:
> But it's still a release candidate.
I agree with you on it will take some years until Python 3.x becomes
widespread.
But Python 3.x is no release candidate any more for quite some time now.
--
Schönen Gruß - Regards
Hartmut Goe
This sounds reasonable :-)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitglied bei http:/
tware quite fast, because all the
backward-compatibility-stuff will bloat the code. *shiver*
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http:
ad to double when
> using
> realms and/or split-DNS.
>
> Double? I don't understand it.
If hostnames resolve differently on arbiter and worker, the worker may
access the wrong machine.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
G
PDF next time?! So
it can be locked at within the mail reader.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
G
o the docs, "alias" is what I would call a
"display name". So "alias" should become "display_name".
BTW: Resolving the hostname on the arbiter may lead to double when using
realms and/or split-DNS.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-In
Am 14.01.2011 10:56, schrieb nap:
> I hope realms and all the distributed features are considered as
> "enterprise grade" features too ;)
Of course. I should have written "another".
But nevertheless please stop full-quoting. Thanks!
--
Schönen Gruß - Regards
Hartmut
ing in it :)
Yes, and send the password unencrypted. *gnaa*
Please stop full-quoting, it's awesome!
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatlich
proposing "enterprise grade" features :-)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mi
lemented, this means: Use it.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitglied be
logs to the broker) and one
local (rotating log file, errors and warnings). If some admin wants more
local logging, he simply changes the config file.
@all: Please do not full quote. Thanks!
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spe
be build on top of the current remote logging
facility. Or can be enhanced to not format the message by it self, but
pass it on the toe remote side to be filtered there.
My idea is to switch to *completely* the standard logging module. No
more print-statements anywhere. Everything is going via loggin
nous communication, too. No need
for backlogs. [Gosh, I'm repeating myself :-)]
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.ci
ndicusses roadmap: Pass messages between
components using a asynchronous mechanism, this is message queuing.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
M
my personal, unpublished and undiscussed roadmap there is a point
"using standard module logging for logging". But this is quite a big
change, so we'll have to discuss it in depth first.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spez
use one sees what went wrong. If the
system makes assumptions, this may lead to wrong interpretation,
inherited errors, more interpretation and chaos. This may destruct
(monitoring) data and lead to wrong alerts.
A small leak will sink a great ship.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-In
stall
So the proper way should be
a) backup your config
b run python setup.py install from the 0.4.1 release
c) restore your config
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www
ut in the log, and not only in the debug.
IC. I should have looked at the source :-)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: htt
FILE
Being in nitpicking mode today :-)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http://www.cissp-gefluester.de/
Goebel Consult mit Mitgli
Am 27.12.2010 14:08, schrieb nap:
> Yes, but shinken manage 2.4 too :) (for RedHat 5 compatibility :( ).
>
Since pysqlite3 is the more enhanced version and is coming with current
Python versions,
I propose using pysqlite3, if this fails try pysqllite2.
--
Schönen Gruß - Regards
Hartmut
boolean text, this is either a programming error
or an error in the config file. Both have to be fixed are the origin
place. Fall-backs are a bad solution IMHO.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexe
his will become ugly code spread
all over the source and quite sure fail on MAC. See
<http://docs.python.org/whatsnew/2.3.html#pep-278-universal-newline-support>
for more information.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist
>
Since this seams to trap others, too, I suggest to add a test for this
to the code.
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP, CSSLP
Goebel Consult
Spezialist für IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de
Monatliche Kolumne: http:
Am 20.12.2010 20:38, schrieb Gerhard Lausser:
> if h.realm.realm_name.strip() != hg.realm.strip():
For me this looks like realm_name should not be a StringProp() (in
post-0.4), but some ListProp() which normalizes the list elements on
reading.
--
Schönen Gruß - Regards
Hartmut Goebel
D
is why :-)
>
> Emails are quite good too :)
Yes, but it's easier to refer to concrete lines of code in the patch. I
have been sceptical on this, but changed my mind. Esp. for larger
changes (like this one) this helps finding bugs or errors early.
> Thanks again for the patches and the
d = True??
I just uploaded a patch for fixing the issues named in the last three
list items to <https://sourceforge.net/apps/trac/shinken/ticket/198>
As you remember I suggested rietveld for discussion patches. This is why :-)
--
Schönen Gruß - Regards
Hartmut Goebel
Dipl.-Informatiker (uni
1 - 100 of 111 matches
Mail list logo
