Bestand wählen
Merken

Fully Public Puppet

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
it was found that the subject there is that any better I'll try to speak of as well right the so we'll an interesting talked the last hour we receive here 1st and stock event and as so I'll quickly review do a quick overview for people who were here for that last part and spend the uh and I both work at HP and I were fully upstream in the OpenStack infrastructure projects on and he works so much and when that happens I'm working downstream in OpenStack uh uh team within HP that consumers are infrastructure and so his talk with largely talking about the direct downstream side of ways that they were consuming our infrastructure and then my talk is talking about the same thing and from the upstream side but with the deeper dive into exactly what we've done with pop it to make it more consumable by other people and so I called the stock on fully public public on because instead of just releasing individual modules to the community and we've open our entire configuration that you have uh example that that works and in production and you can look at how we've set things up and pretty much duplicated and so this is my 1st Boston and very exciting and I've been on a Linux that hobby about 2002 and I got my 1st job as a junior sysadmin where was racking servers and doing on installs Malacanang stuff in 2006 um I'm like drew to going to Delhi in an OpenStack of the big project I worked on an open source and usually from the option packaging science so it went to and I do a few community things and I also host a bunch of servers for community members to use and puts stuff on armored OpenStack and money Infrastructure team which I'll talk about in a moment and then as I say we've we've open-sourced are infrastructure and we have over 100 projects now I to look this morning with 109 and I can bring up of this link this
this here so if you start looking through this list and these are all projects these are all things that were running our infrastructure whatever policies that everything that is used in the OpenStack project has to be an open source projects and partially because we have a huge community of that of contributors come from various giant companies and we don't really wanna step on toes by preferring 1 company's product over another because inevitably you were using a product on the market there probably somehow involved with OpenStack somehow and so we don't really want to put up a barrier between our that the people were consuming of exact and we also always wanted are infrastructure to be an open source project and to be an open source project ourselves we have to use all open source components and so in some cases we've pulled things in how we use Jenkins for continuous integration system and and the Garrett for code review interface and and we've written a bunch of stuff which we open source and also the scroll down emits we have a ton of Puppet modules on which I'll talk about in a moment and these were all act Moseley's reccy recently split out which I'll talk about but
hopes that knowing what
happens that's but I wasn't sure where mechanism
is going food and so the reasons but why why we did this and suspenders talk talked about us having a downstream and and some people actually are consuming a infrastructure so that's an obvious reason why we want do this but more generally and perhaps you're not an open source project are you maybe have the infrastructure internally arm and you're on the off steam and no 1 else ever sees that infrastructure but they're using it every day and you may also be someone who is working on open source project and you put your Puppet modules online but maybe your dog mentation is a great and because a lot of modules part of my routinely download modules and then have to sort of figure out what I what they want their variables like if they want SSH key I have to sort of dig into the manifest and find out whether they won a file or whether they want a just a string or something know that's as the SSH key whether the public or private you have things looking for and I mean it's obvious from looking at the manifest but but then I do have to dig in and find out whether this and so our researcher since we have everything online you can see examples of of what were actually running we found it also encourages better practices if everyone's looking at our stuff and they can be easy as for working up other modules and working on internally for to hard-code everything in and a copy in modules and but to when everyone's looking had people are consuming your infrastructure you end up doing things a better way especially when you're doing code review like we're doing on everything several everything that goes into our repositories peer reviewed and so for us when we finally did the agreed upon a 3 which was actually in September about 2 weeks before 2 . 7 was end of life and even though a really really worried about it the upgrade actually went pretty decently because we are given and doing sort of best practices when it came to public the for us so this is a big 1 allowing others in your organization or in our case all of OpenStack or pretty much all the internet I'd suggest changes and if you're running is inside of a company and you've open-sourced your infrastructure sort of the company you could get repository up somewhere and it allows other people in your organization to understand your infrastructure by looking at what it does and then if they really have obtained point in the workflow they can look at your configuration they consume submit changes or submit more intelligent and suggestions for changing things an organization has allowed developers is really helpful but because the priorities of the OPS team may not fully match the priorities of the developers so the ob steamy not have time to write a Puppet module for the thing the developer wants to deploy but if the developer was taken into themselves they can go ahead and write a module and have Dobbs deployed this works out really well and OpenStack since again there's a lot of companies investing time in OpenStack and money and so they all have their own interests and of course people on the infrastructure team we have our interests as well and the people on the infrastructure team can come from a variety of companies and and it's we usually play favorites as to what's coming in next month's allowing the US the community to submit changes to our infrastructures than super important on vitamins questions registration and the and we're also really hyper committed open-source Nava structure team I'm a lot of us have been continued open source for a very long time and a number of us sort of took this job here because we were sysadmin and we really loved open source and when I my last job was just a regular sysadmin job applying for museums in small medium businesses doing mail servers and web servers all kinds of things and when I went to look for a new job I was like dear Internet I'm this is the man who loves open source is that a job turn that it is the and I got ones so we're all really committed open source and we think sharing is nice so we will we open source the whole infrastructure because we believe in that and we want our infrastructure itself to be an open source project the so now that maybe I've convinced you that this is a good idea and am going to a bit about how we've got about doing this some of this we did sort of from the outset and other things we've had to learn along the way Europe stock based the state of things about 8 months ago was much different than it is today now that we have a very active downstream who's working with us so the 1st thing is is to prepare some policies inside your infrastructure but for us it's easy for us to say we want open source policy everything we use has to be open source and of course that's not really a reasonable thing for a lot of companies and so if that is not the case I'd 1st segregate out the proprietary from your infrastructure and don't open source that part even if it's just configuration files I think the jury is still out on whether you're allowed to share proprietary configuration files from His licenses on those are very unclear on they could be license of the rest of the software the what I but you don't really know when you can ask your vendor and they may be like what does that mean license on a convict file and they may not be able to share and so make sure the proprietary servers outside of whatever your open sourcing and in that year in in a new non violation of any licensing for that the and the open source stuff all easy and find so you can keep that in open source part so licensing convey files and the I I at the full part of models employed for degree because they do have a license but sometimes 1 new find my get hub they don't I've had to e-mail authors of licenses error off authors modules a few times and either because their license is not clear they may mention it offhand in a readme but but that's not actually enough for us to use it because we wanted to be clear and so if Edison attaches to add a license file and any male authors really confirm that this is an open-source project and the don't make me do that it's a waste of all of our times you're putting on get it's probably open source thing and and all of the little thing about licensing convict files but were just admins and that's something we have to start thinking about different on already the this so direct and public side hello 1 things you should be doing generally and is using existing modules from I think we all know this even if we don't necessarily do it offer forking modules we've done our fair share of that um but you should try to use the models that exists and if they're not working for you try work with upstream to make model better and even if you have to fork it you know that happens and we fought the Apache module in our in a big mess and the bicycle modules the and then when you write your own module's this is something we've we've struggled with as well writing them with the intention of of sharing them with others and so I wrote this review day module while back and it should it should be able to use by anyone who's using bugs on launch pad on hooks in and tries to come and is using Garrett as of extended Garrett and and launch pad and tries to determine and so the heat of over the code review so it trying to get developers to hone in on on the reviews that are most important in the queue and so people happen to be using ours hopefully they can use my module and and we share that and and then you wanna keep these outside of your configured like you're general configuration from so early on we had this monolithic thing we called configured and that had a modules directory and and that all of our modules from all the ones that were listed I'm on the stage like this you
know the review day review stats and all these things where you know the public wants yeah oscillate in know planets all review day these were all inside of our monolithic in Figs of someone want to consumer stuff they had download that whole huge thing which included all the stuff that they might not want the but we split the amount and then and then have a server
configuration which I'll talk about in a moment which is here so in our infrastructure 0 we learned that instead of having a single configure we actually want is something that is our infrastructure configure which we've called our system configuration so that's all the definitions for all of our servers in our infrastructure and then had a pretty much duplicate infrastructure server wise and service wise and we then have project configuration which is all over Jenkins job definitions and everything that were sending to Garrett all the OpenStack specific stuff and I can show
you we can see so if I go where system simply cannot but the right so system can figure has has things like our install modules and puppets files and these prepare all of our nodes that we need an and then we have this OpenStack projects module which has manifests for all the servers that we use and so it shows you have a set of cacti and and other things and these are things that other projects consumer naked sort of pick through this list of what they wanna use and and consume in their own infrastructure are probably call them module something other than of 2nd row for the SEC project earlier naming of the project
conveying and then this says again all the OpenStack's really specific stuff and so we have a like Garrett project start yeah MLE the minds of area found that defines all the projects in OpenStack so if you're running your own Garrett you'd wanna you wanna set up a project and think that has all your definitions so again we split the cell from our configuration are system configuration because we don't wanna have no ownership over OpenStack stuff with the thing that we're shipping and all of our system could things with is they're very separate things in the mind of a working open 2nd year near-continuous rich environment that and this is something we did fairly recently and we
actually have a specification for its and which in OpenStack is sort of a plan for hired and do something and so we were in a bad state earlier and then we made a planet that's big long page they can read that I have a link to it later but and went to work items to serve explaining to people if they wanted to split out similar things and this is what we did to do the thing that we did when we got into trouble
um we also had displayed at all of our and non-sensitive custom configurations so ssh keys those are pretty private are host names on our servers we care everyone can see those if they want and so in
hours this is a but an example of our odds zool which is sort of our our queuing thing for 4 tests arms of the patches are the system going to so old that zool cues them up for testing in Jenkins and also process them crosses the resultant Jenkins back into work code review system a so we have things like our status you Alice status that OpenStack that word so is a little bit you know 1st if someone else is working on this project they're probably not going to have that you around because that's hours and and we have all these things these variables you come from higher up so the the the host and the all kinds of different variables that we have here as so these are all split out into and these are all things that are on top of the so using a lot of variables we use variables
everywhere we try not to hard-code anything in common with the we've gotten rid of a lot of the hard coding and now we have a downstream is actually something as patches of saying stop doing that there was 1 actually when I started working on a project that coded . award at the end of everything and so when I was testing it on my own infrastructure at home or host name as a dot com that was of serious pain I don't know why we assume . organ not anything else but it was way it was some enemies until called hirer for all their sensitive data and that's the only thing we put in there and so are ssh keys are passwords and all the user data we have a funny story about how we learn that e-mail addresses are used personal data secret sensitive and once upon a time we said where open source project we posted the open-source mailing lists are e-mail addresses all over the internet so they were not as sensitive data piece we put them our infrastructure code does anyone know what happened the no no the no yes the cities and support request for third-party products so it was sort of related to that it was people downloaded or modules and installed them and then their infrastructure with the e-mailing us all the data up so we knew all about all these infrastructure that were being what's really cool cool there's 1 and you know that company that interesting How underwear that p is coming from its so we decided to save people from themselves food and we took out e-mail addresses and make people do it themselves so that was very fun and then you know cost for support request people joining our channel being like hey you know so and so was working were like we 1 they're running in front cool yeah so but you know there's a sensitive data and public is people download your public models Moses run without reading them this it's kind of scary and some of the really cool things that we've done so we said above a dashboard called public board and we used to use 1 I called public dashboard but it was it was stopped stopping used it was deprecated and that sensor actually came around he showed us by the board which is an interface for showing when when changes are are applied on so I can bring that up here so you
it's a web interface on for us anyone on the Internet can get to it of of the board but OpenStack toward so we had to disable querying because that lets you figure out scary things about our infrastructure that we going to figure out like passwords and so the running this in the public to turn off the query and we've had that you know for all the stuff since were running this all the internet we had to do similar things for other things that we run on but this was a big 1 in public board we had to turn that off and so it is an of how well you can see there's not looks OK but every time every time puppet runs it updates the dashboard reusing and of what a public DB that database backend for profit which motherboard queries so every time a changes done in puppet you can see an update here um so I think I found 1 you yeah this is the Jenkins server and so this 1 actually was an error it skipped a bunch of stuff it failed on something of and order the Jenkins DEB server it supposed to break so in unless puppet run a bunch of stuff was Skipton and failed in and broken but but this allows it allows a so that people were submitting patches and getting their patches merged into a repository these people are random people in the OpenStack project they don't have shell account on our servers they have no idea what's going on they know that the patch was merged they don't know that is picked it up yeah they don't of the service picked it up yet and so by allowing them a peek into into what's going on in the service to this dashboard we make it so that they don't have to ask us that all of this and tell me like it here the get repository 25 minutes ago was showing up and sitting there having refresh a browser to see if the change has taken effect and instead they just go to the dashboard in they can figure out what happened this also you can also dive in add to this a bit more to find out exactly why something failed they give you yes if you click on something and expand and tells you a know could not retrieve the this environment thing so they can they can actually start writing a patch to go ahead and fix what broke what didn't work in the patch without again having to ask us some In for channel on a mailing lists and so really in powers are users who were submitting patches and to do things sort of atomic eponymously no and we don't need to ask us to to sift through log files and that's nice because who likes to do that and so
having having a public thing is is pretty cool on this is a simple the change that was successful yeah lots of free scrolling down through here you can find was the unchanged and changed this is the failed Jenkins the 1 I pulled up um the pretty much was last run and if
you're change was taken effect Ch
and and then and then the final services sharing this and it's great to think you're infrastructures open-source because you put it up on a get repository but you also need to tell people about it and as I mentioned for adolescents to it has then I'd people on it ask you and then in needs documented this is a very important and you cell people where to find its and so the links here to here you get repository get however whatever you using and you wanna give them a workflow of for using the configuration and so
something that we've done is we have yet so
we have a c I don't openstack . org this is where all of our documentation lens and 1 of the things that we do is we tell people exactly how to make a change in public so that's a lot of words but if you wanted to read it is probably good to read them and then is go down here and actually shows you what steps to follow to replicate the standard node in our in our configuration so this is assuming you using Devin Auburn to um use all get you clone our system configuration and then you create your local but pp which would be your proposed change to our system and you run aerosol mop-up puppet and so module scripts that will give the same version of public were using and download all the modules that were using our infrastructure and we give you a simple command you can use a puppet apply command to run your stuff and so then you can test that your change works against the infrastructure that we have and this is been super valuable to us because people are able to write modules and I was writing models I actually I'm about to get shall access to our servers but I don't have it yet so I'm working for 2 years through a code review system were writing cover modules for servers I've never logged into we want to make it possible to be for people because that is the way our infrastructure is built to run giving SSH access to everyone the was project would not be fun at all even if they don't have root itself um document some sort of workflow so people can test the change against the environment you have and then give instructions for how you wanna contribute in our case we use a code review system that OpenStack users so all of our contributions to our infrastructure are done in the same way there for every other OpenStack project and in your case it maybe someone submits a patch on a ticket or maybe on a boat bug reports or are they actually have maybe direct commit access to repository every scary but this some companies trust all the developers molar ops people and maybe that's how it works but to give some instruction on how to do the contribution from for us to think we link that's what 0 yeah we have a
contributing spot here result the Odyssey channel and have dual stuff and all the things here and then explain sort of how
the bootstrapping and glue of your infrastructure works and so how your system is set up and we have a few
workflow things uh where's think with that 1 who no it's not here well apparently I can't find it that's very bad treatment so anyway there there you show people how how the infrastructure goes together so they understand so when the submitting changes to your infrastructure they have they have a good idea of what the arch thing changes against it and and let's that mostly what I had and so in the course of OpenStack open-sourcing our infrastructural another project they're doing this and deviance at their deviance systems administration team they do all the stuff in the open they've been doing it for a really long time I think they've been doing that for longer than most projects I'm aware of having a pretty open team and they have a model that to some other stuff but Mozilla has an engineering team that works on public and they've been open-sourced bunch the stuff and Jenkins infrastructure team there always interested in having more people contribute but so they have an open-source infrastructure came to the 1 of the really great things about these infrastructure teams and OpenStack 1 is that these are real serious project infrastructures that are doing real things in an open source project and if you're a junior sysadmin you're looking for some experience and you can get hired just yet I do not highly recommend volunteering from teams of these seems mostly certain meaning of getting 1 because they don't hire people to the dead in American open-source project but Mozilla Jenkins and OpenStack they have companies who were were funding was the developers and the small enough now but if you start contributing twenties projects 8 I'm I'm thinking of probably get hired so what you want these debates get bigger and word we stop hiring every single person is in that the patch to us and it's still very good real world experience you can definitely put under resonate amen you learned from a real working production of puppet and I hadn't really use puppet a whole lot before I had this job and we were I'm not until you were doing it was horrible now wish we is sent out convey files and packages that we maintain that was theory and so when I started I'd only played the puppet and and then over the course these past couple years I mean i've I was writing my lamp public models of an like 3 months of working with the infrastructure so if you really fast it's it's a great environment to work in and then some resources I I've mentioned throughout the talk from our our documentation of course the repository that has like 109 repositories and novel or open source stuff and these these 2 stacks of again these these are what we did these were the plans for things we're working on and so I showed you the other it the
this out configured 1 um and then there is also a splitting up the public modules and so we actually just finish this on wednesday so super excited be like but we split hours of but models today so quick and we went through me document this we we had this big massive monster and and if you have a similar thing you can come look at our stock and see exactly what we did in some of the cool things we did is we preserved history in the repositories and so if you look at 1 of our puppet can configs here yeah let's go find 1 this is so let's say analysts find 1 good 1 OK let's look at a x a module so even though we just put this module out on Wednesday we were able to use get magic and and actually pull out although although although they get commits we had in our in our old repository and we documented how we did that and say those models new we have a history and get which is pretty cool stuff and so yeah
even though we did things sort of wrong to begin with so now we have specifications to tell other people how also how to get out of that mess um so it is in some ways it's good that we must have we learn a lot um then I thought I had the direct questions thanks again the the the all right so that the question is do we do anything that's close closed-source India was like infrastructure that I am presenting Russia but no I we have a policy if it were but from the OpenStack Foundation is as we cannot use any clothes or sovereign our infrastructure like it would it would it might upset some of the people work injury oversight the company's if we use you know this tool over another tool if we stop paying for that tool and also we do want this to be an open source infrastructure that anyone can replicate the 1 exception we have right now is reason Transifex for translations if anyone knows the story here Transifex used to be an open source project that we're using their hosted version because we that was good enough and but the last last year they went closed-source they no longer update their get have group so for us it's no longer an open source project so were actually switching over to another tool are called Sanader so we had to have a specification of force with my migrating over that but that's that's only because we got stuck in this position where a company owned the source code relicensing they realize they all the changes were word up to a certain point they just stop maintaining their open source code so not intentional that and allied why do we use public not something else and we had a really really small team when we started an uh the guy who is implementing and he tried to get stuff running ending up of running 1st and it's I I I will say I mean by that is that that is fine if we're starting over with values sensible because we don't 100 play clients everywhere but were already pretty much and you know if it if it now we only have them everywhere we look as this is already there we already have that because as a result talk to each other but it it works and that there was we don't we don't pay any meaningful evidence configuration management things specifically and this and I think a lot of lessons here could be used for other configuration management systems as well so yeah the all of the the of the what the right so the question was uh if you're smaller projects open is huge we we have like 8 100 test servers in our CI how how reasonable is 4 per smaller project to consume some of this and it's actually quite easy um In inasmuch as infrastructures every easy and continuous innovation and we a lot of the things are very probable so we have a uh a flow that goes Garrett zool Jenkins & that is we roads will because we have a ridiculous number of patches like so many factors come through actually look at it the weekend so all be quite so busy but all these are these are timestamp someone changes were updated so the the only updated by every 2 minutes or so this is pretty slow so day and we we have we have superactive patch activity and so we had a right dual because patches would come in and they conflict with each other and they'd hit the repository may have complex if you have a much smaller project for the patch size you know maybe are coming in so fast there's a Garrett all plug-in and they can be used and we've actually documented that in some places where you can just use that instead of using a big school thing I'm so that's 1 of the scaling things that we've done and you can also were using Jenkins with a tool called Note pulled out which manages our fleet of 800 servers that are being used for travel all tests that you could just use Jenkins in old model just having a plug directly into a fleet of service you manage some other way and so it's only reasonable to use different things that we've really done things to scale up on but to use different this it's all pretty portable from and for things like our status spots in our our our upcoming bug tracker that word developing um those are all they can be run in small batches the the type of thinking and the
Offene Menge
Momentenproblem
Systemverwaltung
Open Source
Systemverwaltung
Binder <Informatik>
Biprodukt
Modul
Ereignishorizont
Keller <Informatik>
Richtung
Konfiguration <Informatik>
Systemprogrammierung
Open Source
Menge
Prozess <Informatik>
Server
Projektive Ebene
Installation <Informatik>
Konfigurationsraum
Momentenproblem
Open Source
Kontinuierliche Integration
Spieltheorie
Mailing-Liste
Physikalisches System
Biprodukt
Modul
Weltformel
Code
Open Source
Zusammenhängender Graph
Projektive Ebene
Feuchteleitung
Schnittstelle
Offene Menge
Bit
Punkt
Gemeinsamer Speicher
Hinterlegungsverfahren <Kryptologie>
Eins
Internetworking
Regulärer Graph
Prozess <Informatik>
E-Mail
Figurierte Zahl
Metropolitan area network
Kraftfahrzeugmechatroniker
Nichtlinearer Operator
Dokumentenserver
Software
Registrierung <Bildverarbeitung>
Server
Ablöseblase
Projektive Ebene
Verzeichnisdienst
Schlüsselverwaltung
Zeichenkette
Aggregatzustand
Fehlermeldung
Varietät <Mathematik>
Sichtbarkeitsverfahren
Hypercube
Selbst organisierendes System
Mathematisierung
Zahlenbereich
E-Mail
Code
Open Source
Selbst organisierendes System
Informationsmodellierung
Benutzerbeteiligung
Lesezeichen <Internet>
Modul <Datentyp>
Software
Vererbungshierarchie
Warteschlange
Datenstruktur
Softwareentwickler
Maßerweiterung
Konfigurationsraum
Modul
Graphiktablett
Autorisierung
Videospiel
Open Source
Konfigurationsraum
Systemverwaltung
Elektronische Publikation
Modul
Quick-Sort
Keller <Informatik>
Programmfehler
Minimalgrad
Mereologie
Offene Menge
Open Source
ATM
Client
Server
Statistische Analyse
Programmbibliothek
Pendelschwingung
Keller <Informatik>
Modul
Umwandlungsenthalpie
Server
Momentenproblem
Konfigurationsraum
Programmschema
Mailing-Liste
Physikalisches System
Objektklasse
Elektronische Publikation
Modul
Quick-Sort
Open Source
Physikalisches System
Datensatz
Dienst <Informatik>
Knotenmenge
Modul <Datentyp>
Menge
Rechter Winkel
Prozess <Informatik>
Server
Projektive Ebene
Installation <Informatik>
Konfigurationsraum
Umwandlungsenthalpie
Offene Menge
Ablöseblase
Automatische Handlungsplanung
Mathematisierung
Indexberechnung
Zellularer Automat
Physikalisches System
Binder <Informatik>
Quick-Sort
Keller <Informatik>
Homepage
Systemprogrammierung
Open Source
Modul <Datentyp>
Flächeninhalt
Offene Menge
Zellularer Automat
Computersicherheit
Projektive Ebene
Versionsverwaltung
Konfigurationsraum
Programmierumgebung
Aggregatzustand
Softwaretest
Offene Menge
Server
Subtraktion
Bit
Physikalisches System
Objektklasse
Variable
Code
Quick-Sort
Keller <Informatik>
Physikalisches System
Open Source
Patch <Software>
Variable
Server
Projektive Ebene
Wort <Informatik>
Schlüsselverwaltung
Brennen <Datenverarbeitung>
Modul
Bit
Sensitivitätsanalyse
Massenspeicher
Nabel <Mathematik>
Selbst organisierendes System
Browser
Adressraum
Mathematisierung
Hauptplatine
Whiteboard
Code
Internetworking
Physikalisches System
Open Source
Informationsmodellierung
Front-End <Software>
COM
Passwort
Ereignishorizont
E-Mail
Modul
Leistung <Physik>
Schnittstelle
Soundverarbeitung
Benutzeroberfläche
Dokumentenserver
Datenlogger
Datenhaltung
Open Source
Machsches Prinzip
Ähnlichkeitsgeometrie
Mailing-Liste
Biprodukt
Modul
Nabel <Mathematik>
Variable
Quick-Sort
Patch <Software>
Dienst <Informatik>
Codierung
Server
Projektive Ebene
Programmierumgebung
Schlüsselverwaltung
Expandierender Graph
Fehlermeldung
Soundverarbeitung
Open Source
Offene Menge
Explosion <Stochastik>
Verschlingung
Mathematisierung
Versionsverwaltung
Ereignishorizont
Keller <Informatik>
Open Source
Dienst <Informatik>
Gerichtete Menge
Verschlingung
Open Source
Konfigurationsraum
Zellularer Automat
Datenfluss
Binder <Informatik>
Konfigurationsraum
Nabel <Mathematik>
Lokales Netz
Offene Menge
Mathematisierung
Code
Überlagerung <Mathematik>
Open Source
Knotenmenge
Informationsmodellierung
Skript <Programm>
Wurzel <Mathematik>
Softwareentwickler
Konfigurationsraum
Gerichtete Menge
Konfigurationsraum
Mathematisierung
Physikalisches System
Modul
Quick-Sort
Programmfehler
Keller <Informatik>
Patch <Software>
Verschlingung
Server
Kommensurabilität
Projektive Ebene
Wort <Informatik>
Datenfluss
Programmierumgebung
Verkehrsinformation
Offene Menge
Dokumentenserver
Bootstrap-Aggregation
Mathematisierung
Automatische Handlungsplanung
Keller <Informatik>
Extrempunkt
Physikalische Theorie
Open Source
Informationsmodellierung
Prozess <Informatik>
Reelle Zahl
Elektronischer Fingerabdruck
Softwareentwickler
Dokumentenserver
Amenable Gruppe
Open Source
Güte der Anpassung
Systemverwaltung
Physikalisches System
Biprodukt
Elektronische Publikation
Quick-Sort
Keller <Informatik>
Arithmetisches Mittel
Patch <Software>
Projektive Ebene
Wort <Informatik>
Programmierumgebung
Offene Menge
Server
Dokumentenserver
Ablöseblase
Mathematisierung
Benutzeroberfläche
Modul
Keller <Informatik>
Open Source
Informationsmodellierung
Heegaard-Zerlegung
Parametersystem
Modul
Resultante
Offene Menge
Server
Punkt
Ortsoperator
Dokumentenserver
Mathematisierung
Versionsverwaltung
Gruppenkeim
Abgeschlossene Menge
Zahlenbereich
Hausdorff-Raum
E-Mail
Code
Richtung
Hypermedia
Open Source
Informationsmodellierung
Client
Migration <Informatik>
Datentyp
Translation <Mathematik>
Zeitstempel
Modul
Softwaretest
Umwandlungsenthalpie
Dokumentenserver
Textbaustein
Open Source
Ausnahmebehandlung
Physikalisches System
Quellcode
Datenfluss
Quick-Sort
Teilbarkeit
Keller <Informatik>
Programmfehler
Patch <Software>
Konfigurationsverwaltung
Dienst <Informatik>
Forcing
Rechter Winkel
Server
Projektive Ebene
Wort <Informatik>
Stapelverarbeitung

Metadaten

Formale Metadaten

Titel Fully Public Puppet
Untertitel The open source OpenStack project infrastructure
Alternativer Titel Configuration Management - Public Puppet
Serientitel FOSDEM 2015
Autor Joseph, Elizabeth K.
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 10.5446/34340
Herausgeber FOSDEM VZW
Erscheinungsjahr 2016
Sprache Englisch
Produktionsjahr 2015

Inhaltliche Metadaten

Fachgebiet Informatik

Ähnliche Filme

Loading...
Feedback