Bestand wählen
Merken

python-netsnmpagent -- Writing net-snmp AgentX subagents in Python

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
thank you and I and so this talk will do what so we know that the present location that SNP it's some of you probably already knows is the dominant toolkits to as a monkey on on Unix so that's basically it's it's applications so some of you work longer applications mostly and the libraries of the status of a mass in the 6 and the more you can learn what models and see which of them include viable the DLO consisting call has got some other extension mechanisms as well as more the genetics and particularly some species of CAP loving scattered standard generator matrices so you can know if you if you focus on so you can just given it but it will you would just call me to see unattended were generated as this was a template for your own age so you don't have to to to write everything itself just for the missing pieces that work but it's see obviously and what they do not exclude its use only beyond this talk so while I think by expressed will be so playing simple and so I could be as well but so it can hide from the net as the count of a Python module that doesn't the dope and it's quite a large pile of of of C code so that affects the old CAP API into into an object so oriented approach but it's a client code only so if you want to do an agent the stock on so basically the the idea is some we just take the CAP API and we excessive from Python with the types um imitating eugenics subregions with Nancy no matter what the first one to have you there's some models sourceforge called by genetics so but it had some design issues and it looks whether offered to me my did some pictures initially but after seeing them not being considered for for for weeks and months that I just thought well the the whole shouldn't we by myself so I understand which is of course an attitude to which is well known among developers just to do so hello from and the agents but it's pretty simple actually just to source files no 1 no 1 has all the C type stuff which means to for the types as we would say the around which works the API called just think about and then move on top of its list that isn't the agent PI which is still in the fight 4 classes we see about that but it is also not not not quite large but this extensively commented which explains why the wheel conciseness one-third less on other got example of a recent agents included what mean coding style and it's not particularly PAP 8 or whatever the price coding style and proposal is small aswell encoding side that is consistent it's been tested move on both 5 . 4 and 5 . 7 versions of medicine the for the simple reason that the thing it was targeted at enterprise these distributions as they call themselves Sousa and to a bunch of and they're still shipping 5 1 4 so I have to take care of those as well let's look at some Hollywood's
implements a lot might the now this is basically the mean vectors and rotation in how you construct levies just give it a name on just have the data type in this case it's an unsigned 32 bits of scaleable you the following make successes that was not so important what are we give a description in terms given the referring somewhere in the S and P M beach we you what it this subject to appeal so this simple skills 1 translates to something like don't want undefeated takes a set of this is a quite simple example it's just an integer value which we want you to to make available and you could have been can also tables but that's way more complex that is a part of interest so if you have a model like this this is normally decoration just saying on the just Milliken slightly to read for information I mean you don't have the code yet and that's of course the intent of using my Python Edison pH model so you want to park and go to the prom impulse they didn't just in part 2 1 5 of the class nor the API calls you construct a new objects of type is the agent which needs an agent name which that an appeal we cry isn't only on any given a list of firm that flies in this case the mid file we just defined in that is NP requires that some for example so it can translate into only between them symbolic names in the numeric or these we've just seen so you quite an agent object constant public sex exception handling unlike you next thing you do know just because you've come take away the structure in the in the final and doesn't mean that all those some information elements are have been appearing medical leave you still have to explicitly which is them which makes sense because you could have 2 agents implementing different parts of 1 the 5 so you come other than we're exploration some of up sort of a cost factor really media for the different data types in this case integer 32 thought was on-site white figure for whatever they mean any given the the already with this into 2 and should be wettest which below so inside here my between once you've done that you interested all your own a snappy objects I'm speaking of SNP objects because you could just as well which is still a table is basically woes themselves of simple color values and once you've done that you call agents or attack and that makes it actually connected loss regions as in the to consider proper exception handling it again and then the looks like
this 1 is a simple example of course is up to an endless loop room call agent checking process all the time this is in fact and that is in the lab we call them they designed this way and as I said I'm just working the API calls and so we have to do this way as well the this example I could serve for from them update the integer value just declared and publish all the time provided that would have newish initialized that we get this whatever and um I could call this the piece that's from the command line elements in the new value of here and this is very simple there are also more complete examples and for some of tables in the source distribution and we have taken this course also much more complex so the agent of vote for internal users like 100 kilobytes 3 thousand lines so because you want to do all the things you stuff like demonizing and and and look for mapping and whatever what
I mean if I have time to ensure optimal in in a 2nd just so you know that something's missing of course modifications entrapped in particular so you can as American white an agent which informs the mass agent and for example for thank you anything like that API documentation who would have thought that and to maybe some unit tests because as I said after all we produce safety white so it would be of course nice to have the elementary is optimal use of a safety at all going to give a real
short trip tomorrow this the examples
directory of the source distribution kind of got the simple agent you have a look inside this is all things so we just saw basically scholar values different data types and also tables and the nice thing is you just 1 script it's it's all there were things necessary to test all things like them calling isn't PDE in it is isolated from the west of the system so it can actually wanted with a system-wide has impeded instance 1 unit of the problem the commodity of 2 1 an advance the current values of all those the variables and so I can go out from separate
so just give the command like like you take this new fire from this we connect to the instance on this port because this a set of separate from system-wide PD and adjusting the values and I can do things like you know the like ICA also success the use of the tables if I had to wait till the natural state of what so in
this example it's some room should be 10 which is
so anything like that and they could play while with SNP set as well but because sometimes up and questions our word on sort of all in terms of what's in terms of the yeah the question was and how much overhead some of the patent annotation would be compared to see that in in what terms mean and like like the ones speeder work all here the rest of world and the real time in the the core of the of the it's basically so it's of where it's very close associate the I so if you if you have ever delve into the details of the CAT II will recognize so much of it and then again this is the it's free of just given aid grade an objects and practices some variables and you got a message from the all that and benchmarked it's them for simple reason that 1 means the same in the same agents I write them was options so I have moved comparison but some of all of them I haven't seen any performance issues so far because Python is actually quite fast I'm running out of time to the next or pesticides and thank you he does got not therefore what and how the questions and questions answers you know I'm not of the room and the next so was started in a couple minutes the the new
Distributionstheorie
Gewicht <Mathematik>
Klasse <Mathematik>
Versionsverwaltung
Kartesische Koordinaten
Euler-Winkel
Zählen
Template
Code
Computeranimation
Open Source
Mailing-Liste
Client
Informationsmodellierung
Modul <Datentyp>
Code
Datentyp
Programmbibliothek
Maßerweiterung
Softwareentwickler
ART-Netz
Modul
Kraftfahrzeugmechatroniker
Architektur <Informatik>
Matrizenring
Benutzerschnittstellenverwaltungssystem
Codierungstheorie
Streuung
Template
Systemaufruf
Ruhmasse
Mailing-Liste
Quellcode
Objektklasse
Natürliche Sprache
Modul
Kugelkappe
Objekt <Kategorie>
Client
Programmierstil
URL
Versionsverwaltung
Unternehmensarchitektur
Standardabweichung
PROM
Distributionstheorie
Abstimmung <Frequenz>
Subtraktion
Einfügungsdämpfung
Prozess <Physik>
Tiefensuche
Klasse <Mathematik>
Sprachsynthese
Drehung
Element <Mathematik>
Term
Code
Open Source
Loop
Deskriptive Statistik
Informationsmodellierung
Datentyp
Translation <Mathematik>
Datenstruktur
Figurierte Zahl
Gerade
Modul
Tabelle <Informatik>
Einfach zusammenhängender Raum
Antwortfunktion
Systemaufruf
Ausnahmebehandlung
Mailing-Liste
Quellcode
Vektorraum
Lineares Funktional
Gerade
Teilbarkeit
Quick-Sort
Dialekt
Mapping <Computergraphik>
Objekt <Kategorie>
Arithmetisches Mittel
MIDI <Musikelektronik>
Konstante
Menge
Komplex <Algebra>
Ganze Zahl
Hypermedia
Mereologie
Kantenfärbung
Information
Registrierung <Bildverarbeitung>
Tabelle <Informatik>
Softwaretest
Einheit <Mathematik>
Existenzaussage
Einheit <Mathematik>
SALEM <Programm>
Ruhmasse
Trennungsaxiom
Distributionstheorie
Subtraktion
Hyperbolischer Differentialoperator
Physikalisches System
Quellcode
Einheit <Mathematik>
Menge
Datentyp
Skript <Programm>
Verzeichnisdienst
Aggregatzustand
Tabelle <Informatik>
Instantiierung
Freeware
Paarvergleich
Term
Quick-Sort
Computeranimation
Eins
Konfiguration <Informatik>
Gradient
Objekt <Kategorie>
Variable
Echtzeitsystem
Speicherabzug
Wort <Informatik>
Overhead <Kommunikationstechnik>
Message-Passing

Metadaten

Formale Metadaten

Titel python-netsnmpagent -- Writing net-snmp AgentX subagents in Python
Untertitel Implementing custom MIBs made easy
Serientitel FOSDEM 2014
Autor Hollants, Pieter
Lizenz CC-Namensnennung 2.0 Belgien:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen.
DOI 10.5446/32611
Herausgeber FOSDEM VZW
Erscheinungsjahr 2014
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract python-netsnmpagent is a Python module that facilitates writing Net-SNMP subagents in Python. Subagents connect to a locally running Master agent (snmpd) over a Unix domain socket (eg. "/var/run/agentx/master") and using the AgentX protocol (RFC2747). They implement custom Management Information Base (MIB) modules that extend the local node's MIB tree. Usually, this requires writing a MIB as well, ie. a text file that specifies the structure, names and data types of the information within the MIB module. This lightning talk will give a really quick introduction to SNMP and MIBs and show how easy it is to implement your own custom MIBs using Python and python-netsnmpagent

Ähnliche Filme

Loading...
Feedback