How To Build an Openfire Chat Server On Debian 5

palegray.net writes "Inspired by a recent Ask Slashdot, I've written a step-by-step tutorial for setting up an Openfire server on Debian Linux, for those interested in running their own open source collaboration server. Aimed at those just getting started with collaboration software, the tutorial shows precisely how to get Openfire up and running quickly on a base Debian install, and offers a basic feature tour of the software's plugin and IM gateway functionality."

Comments

How To's are so 90s..

[Anonymous Coward]

These days you make me a VMWare image I can just pop-in and run.

Re:How To's are so 90s..

[GMFTatsujin]

Sorry, that's out of my league. Do you have a how-to on using a vmware image?

Re:How To's are so 90s..

[Pollardito]

no, but I have a vmware image of his vmware image you can just pop-in and run

Re:How To's are so 90s..

[jnetsurfer]

I love Virtual Appliances. But HOWTOS are still necessary because I, for one, always question the security/authenticity of 3rd party VMware images. I use them to eval software and, if I like it, create my own appliance so I have full control and know there are no rootkits/backdoors installed...

Re:How To's are so 90s..

[fm6]

That's fine for a demo. I wouldn't want to run a chat server with hundreds of users that way.

Re:How To's are so 90s..

[99BottlesOfBeerInMyF]

That's fine for a demo. I wouldn't want to run a chat server with hundreds of users that way.

Do you mean you wouldn't run a production server as a VM image or you wouldn't run it using a VM image you did not create yourself? If the former, you should know a lot of servers are moving to running VMs for ease of: scaling, hardware sharing, backups, hardware migration, and cheaper "in the cloud" hosting. Why wouldn't you want to run your production server in a VMWare VM?

Re:How To's are so 90s..

[rufus t firefly]

For the lazy people here, try this one: http://www.vmware.com/appliances/directory/64572

Re:How To's are so 90s..

[fm6]

Do you mean you wouldn't run a production server as a VM image or you wouldn't run it using a VM image you did not create yourself?

The latter. Recall that TPP thought that VM appliances made installation how-tos obsolete. That was the idea I was pooh-poohing.

As it happens, I'm the documentation lead for a server for which virtualization is very important. A very common use case for this system is "consolidation": you take a rack full of low-end servers each handling a separate application, and convert them all into a single machine running each application in a separate VM. Savings in power and real-estate cost can be enormous. VMware support is a key feature (of course), but Sun is also pushing its own Xen-based solution.

Re:How To's are so 90s..

[daybot]

Sup dawg, I heard u like vmware, so I put a virtual machine in ur virtual machine so u can virtually virtualise virtualisation.

Re:How To's are so 90s..

[palegray.net]

You hit the nail right on the head. I do create VMware and Virtualbox images for various appliance setups, and I share them with friends who know me personally. However, I completely understand admins who wouldn't run something I created in their enterprise, as they have no real assurance that I haven't put something nasty in there (of course I haven't, but they're doing their job by asking questions like that).

Re:How To's are so 90s..

[thePowerOfGrayskull]

This is something that I'm not understanding about virtualization.

If you're running 30 different services on 30 different machines, wouldn't it make as much sense to consolidate them to 15 services on each of two modern machines, instead of needing to maintain 30 different virtual machines? That way, you're also getting native performance without the overhead of virtualization.

On the other hand, in cases where someone is hosting thousands of the same application on individual servers, it makes a lot more sense.

Re:How To's are so 90s..

[_Hellfire_]

If you're running 30 different services on 30 different machines, wouldn't it make as much sense to consolidate them to 15 services on each of two modern machines, instead of needing to maintain 30 different virtual machines?

That's something I've always wondered as well. With a virtualised environment, you'd have to keep up the security patches for 30 different environments (among other tasks you'd need to replicate 30 times). Modern OSes *are* capable of running more than one process - stick em all on the same physical machine and be done with it (unless there's something I'm missing).

Re:How To's are so 90s..

[palegray.net]

Security is a bit part of the equation. If one service gets exploited, you don't lose your whole production environment to some kid in Ukraine who got lucky with a zero-day exploit.

Security patching can be (and usually is) largely automated, with mail going to folks who monitor patch cycles on the network.

Re:How To's are so 90s..

[s0l1dsnak3123]

I LOVE Virtualbox. I upgraded it just yesterday when I (clean) upgraded my system from Ubuntu Intrepid to Jaunty Beta. I installed virtualbox, copied the old config directory, it said "virtualbox has detected some old images blah blah blah, do you want to upgrade the files", I clicked yes, in less than a second I was running the image. Job done.

So as you can see, Virtualbox seems to be very good at ensuring that the images will still work even after an upgrade.

Re:How To's are so 90s..

[fm6]

If you're running 30 different services on 30 different machines, wouldn't it make as much sense to consolidate them to 15 services on each of two modern machines, instead of needing to maintain 30 different virtual machines?

I asked pretty much the same question when I started working with this hardware.

In many cases, you could put all 30 in one machine. The machine I described is advertised as supporting 80 virtual machines at once. (This number will go up drastically when the Istanbul/Hypertransport 3 version is available.) If it can support 80 VMs, it can certainly support 30 services!

Thing is, IT people don't like to put 30 services on one system. They don't even like to put 2. When a system goes down, they don't want more than one service to go down with it. Hence the huge market for 1U machines that mostly end up running a single service.

If you're running a half dozen services, spending a few extra K to put each service on a separate server is a no-brainer — you can easily lose that much if, say, a database server goes down and takes your email, web server, etc., with it. But when your IT needs start getting big, that adds up. Which is why VMware is a multi-billion-dollar company.

Fluff Articles?

[grimw]

Something this basic requires a posting on the front of Slashdot? Great choice, kdawson!

I'll be sure to try and get my article about setting Openfire up on FreeBSD here soon..........

Re:Fluff Articles?

[palegray.net]

This was a direct response to an Ask Slashdot article a few days ago. The community seemed more than interested to offer tips for what software they advised using, but nobody seemed to offer any concrete installation instructions. The Openfire community documentation is great, but kinda hard to wrap your brain around for Debian installs (mostly because it doesn't cover prerequisites and some other points).

very odd

[johnjones]

I thought that openfire was a jabber server so when the article states

"XMPP is also quite stable, at least for my purposes."

I get really worried about how much the author actually knows about the server...

realistically I would like to see some mobile jabber clients for things like blackberry

if anyone knows a free beer version of a jabber client for blackberry let me know
(I am not interested in webapp's since I like my privacy)

regards

John Jones

   

Re:very odd

[ianfs]

FYI - XMPP is the Jabber protocol.

Re:very odd

[fuzzyfuzzyfungus]

I think that that was grandparent's point: To see somebody write an article on setting up a jabber server, then casually note, near the end, that XMPP seems to work OK for him, as though it were some sort of esoteric peripheral configuration option, tends to cause one's eyebrow to start migrating up the forehead.

Re:very odd

[janeuner]

if anyone knows a free beer version of a jabber client for blackberry let me know

If you had a WinMo or Symbian phone, I could recommend a half dozen different apps. CrapBerry apps? Not so much.

Re:very odd

[Arkem Beta]

Once again, that section is referring to the inter-server communications.

The author is saying that even though Openfire lists XMPP server to server communication as experimental it has worked fine for him.

Re:very odd

[palegray.net]

I get really worried about how much the author actually knows about the server...

Hi, I'm the guy that wrote the article. You'll note in the screenshots that under the gateway settings section XMPP is lited as "experimental." I thought it was pretty odd, too. That's why I specifically mentioned the fact that it worked great for me.

Re:very odd

[palegray.net]

See my reply to the GP here. Thanks :).

Re:very odd

[fuzzyfuzzyfungus]

Fair enough. I'm sorry about the misinterpretation.

I wonder if, perhaps, that is an artefact of the fact that XMPP is, strictly speaking, derived from jabber, rather than just another name for it. Possibly there are some differences between earlier "jabber" implementations and later "XMPP, informally know as jabber" implementations.

Re:very odd

[palegray.net]

That's okay, I went ahead and put a notation in the tutorial to prevent further confusion on the matter :).

Re:very odd

[Wodin]

Either that, or it could just mean that Openfire is more focussed on the c2s part of XMPP and what is experimental is the s2s functionality. i.e. the "XMPP" they are calling experimental is under "Experimental Gateways" in the configuration.

What about a better conferencing solution?

[Radhruin]

I work at a small business with 10 or so employees. Recently, people have been getting more and more used to Instant Messaging as a way to provide non-intrusive information that is more instant than email. Lately we've even taken to setting up chat rooms to bring together three or four stakeholders to have a short conversation about something.

Now, I know XMPP and OpenFire support Multi-User chats, but what about more robust conferencing? The other day, I wanted to send a screenshot of an application I was working on to everyone in the MU chat. From what I could tell, this is not possible in OpenFire, and perhaps not supported in XMPP. Also, it would be great to collaborate on or point to a file that exists in our shared filesystem, which I would think is a fairly common use case, but I could not find a way to do that either.

So, I suppose what I'm wondering is, are there any solutions similar to Openfire but provide more robust conferencing? It'd be killer to be able to toss revisions around and maybe do some whiteboarding or something...

And if not, who wants to help me write an XEP that will address these use cases? ;)

Re:What about a better conferencing solution?

[jledhead]

openfire whiteboard - http://www.version2software.com/blog/2007/03/whiteboard-sparkplug-released.html openfire more robust collaboration - maybe this http://www.igniterealtime.org/community/docs/DOC-1518 we use openfire+spark with SSO

Re:What about a better conferencing solution?

[Anonymous Coward]

"would be great to collaborate on or point to a file"

It sounds like you should use a URL.

Re:What about a better conferencing solution?

[windsurfer619]

It's actually really easy to code. I wrote a multicast file transfer application for distributing games at LAN parties, and it only took me a day or so of perl hacking.

Re:What about a better conferencing solution?

[gduquette]

"The other day, I wanted to send a screenshot of an application I was working on to everyone in the MU chat"

The Spark client that the same company offers has the ability to send screen-shots, actually you grab any part of the screen you want and send it. That is the one feature that is used a lot where I work.

Re:What about a better conferencing solution?

[Radhruin]

Right, but unless I'm mistaken, there is no way to send a screenshot to all of the people in a multi-user chat without initiating individual file transfers. Collaboration between two people seems fairly well supported, especially when you include the above-mentioned whiteboard application, but nothing I've found works when you want to collaborate among three or more people.

Re:What about a better conferencing solution?

[gduquette]

Oh good point.. I will have to check that. You are likely right.

Re:What about a better conferencing solution?

[Jeremy Visser]

Not if you're in a locked down enterprise that doesn't allow you to just willy-nilly fire up a web server.

Re:What about a better conferencing solution?

[ion.simon.c]

Every enterprise that I know of has a web browser installed.
Every web browser that I know of accepts file:// URLs.

Alternatively, if the OP is in a Windows environment, he can pass around SMB URLs to all participants.

Re:What about a better conferencing solution?

[ion.simon.c]

What prevented you from passing around a file:// or CIFS (\\server\sharename) URL to the participants?

Re:What about a better conferencing solution?

[Pastis]

You could use an extra service, such as Dropbox, and place your files there, for them to download them via the $service web interface.

Not completely user friendly though.

What's openfire?

[zx-15]

I did RTFA, it does explain how to install it, but besides that what does openfire actually do? Jabber support, graphs showing who is online, what else?

Re:What's openfire?

[99BottlesOfBeerInMyF]

I did RTFA, it does explain how to install it, but besides that what does openfire actually do? Jabber support, graphs showing who is online, what else?

It's a Jabber server. What else do you expect it to do? Your comment is like saying you know Apache supports serving Web pages, but what else does it do?

Finally...

[smooth wombat]

someone who knows how to document a procedure. I don't use linux but even I could follow the instructions.

This is one of the key reasons for a slow adoption rate among interested users. Instead of getting the usual, "RTFM newb!", if there was more explicit documentation such as this that people could be pointed to, people would not be so readily turned off.

And no, MAN pages do not count as documentation. Some people (dare I say most?) need step-by-step instructions on how to do something the first time so they are sure they are doing things correctly. Afterwards, they're free to tinker til their hearts delight.

*gives a mold-friendly thumbs up*

Re:Finally...

[Samschnooks]

And no, MAN pages do not count as documentation. Some people (dare I say most?) need step-by-step instructions on how to do something the first time so they are sure they are doing things correctly. Afterwards, they're free to tinker til their hearts delight.

*gives a mold-friendly thumbs up*

Damn straight about the Man Pages! Man pages are for reference for experienced users. In other words, for those who are wondering, "What are the parameters again?" Why?

You see, when you read the man pages, parameters have all these "[]" and "-" and other things. I've occasionally come across programs that had "optional" parameters in their man pages that ended being required.

Re:Finally...

[value_added]

And no, MAN pages do not count as documentation.

I appreciate the general sentiments in your post, but the above is simply not correct. Putting aside issues of writing style, quality, completeness, etc., along with the abomination called info pages, manpages ARE documentation.

What you are looking for or expecting is a Tutorial. And most likely one written in a friendly style that includes a breadth of related topics (like "How do I install this thing?" or "What's a protocol?" or "How does XMMP work?". For that, I'd suggest a Google search. Programmers are expected to document their work, but it's unreasonable to expect them to write Tutorials.

To put things in a reverse perspective, the frustrating thing for non-Windows users is that on Windows, almost nothing is documented. Lots recursive clicky links that pop up HowTo's with explanatory or descriptive verbage, but no real documentation.

Re:Finally...

[MasterOfMagic]

This is one of the issues that pisses me off as a former open source programmer who gave up on it because of user whining.

When you're willing to spare the time to help a project write it or pay someone in that project to write it, then you can complain about lack of documentation. A programmer is a busy person - they usually have a full-time job to pay the bills (assuming their full-time job isn't to work on the project, in which case, you are completely justified in castigating them for a lack of documentation) and they work on free software in their free time. Often, they're more concerned with fixing bugs and working on the next addition than they are about if new people are using their software.

What I'm saying is that most programmers write the documentation after they're done coding, and if they're adding new features (often at the request of the same users complaining about lack of documentation), this occasionally falls by the wayside. One of the upsides of free software is that someone else can come along and help with the documentation.

Or the users can whine and the situation will never get better, programmers will walk away, and people still won't read the (better) documentation. Kudos to the walkthrough writer above for doing it right.

Re:Finally...

[fm6]

Man pages are (or are meant to be) reference documentation. Reference docs are important, but yeah, so are procedural docs.

Thing is, they're really hard to write well. I earn an absurd amount of money because I'm good at writing them. Most developers are not. Even the ones who are good writers (a distinct minority) tend to be good at academic-style writing, where you assume a lot of sophistication on the part of your reader. Such writers have no patience for the nit-picky detail good procedural docs need.

(I once wrote software release notes which included an explanation of how to unzip an archive. The developers, who happened to be pretty brilliant computer scientists doing cutting-edge work, thought their audience was "smarter" than that and made me take it out. Wackiness ensued.)

If you want good procedural documentation, you have to put a lot of work into creating it. Most open source projects don't have the extra bandwidth, and even if they do, they don't have people with the right skill set. That boils down to fluency with language and and stubborn patience when describing boring details. Not a big skill set, but one very few people seem to have.

Re:Finally...

[SCHecklerX]

Um...

A quick google of "linux howto" would take you to this:

http://tldp.org/

Not quite sure what is so difficult about those. Many of the things I now do as part of a career I learned from that site.

Re:Finally...

[onkelringnes]

I think you traverse the wrong corners of internet, cause my impression is that most people are very helpful, especially to newbies. I've written lots of question on a number of different forums on the net and have never been brushed off.

OTOH, if you try to ask a question to a mailing list, you really need to take some time to understand the tone of the conversation there before posting anything. Try asking how to setup your webcam on the linux kernel mailing list....

BTW; HowtoForge-linux is the place to go...

Re:Finally...

[smooth wombat]

I earn an absurd amount of money because I'm good at writing them.

Your company hiring? Part of my job is doing exactly what the author of this posting does (and apparently you). Document, step-by-step, how to do something. I'm writing for people who are supposed to know this stuff already yet I still provide screenshots and exact steps (Click Next, Click Next, Click Next, Click OK) so there is no misunderstanding or confusion as to what needs to be done.

The only time things don't work out as expected is when someone didn't follow exactly what I told them do. Maybe they thought they knew what they were doing and tried to take a shortcut or inadvertently missed a step (I'm guilty of that). Maybe they were just lazy and didn't bother to read (they are government workers after all). Whatever the case, 99.9% of the time there is a problem, it's because the steps I outlined were not followed.

Re:Finally...

[palegray.net]

Although I understand that point of view, I actually find it harder to keep track of where I am in a procedure if it's all on one very long page. I frequently scroll around a bit to double-check things, and it's a lot of easier to avoid getting lost with pages only about 2-3 screens high :).

Re:Finally...

[Anonymous Coward]

"What I'm saying is that most programmers write the documentation after they're done coding"

They are bad programmers then. Full stop.

Re:Finally...

[Just Some Guy]

Damn straight about the Man Pages! Man pages are for reference for experienced users. In other words, for those who are wondering, "What are the parameters again?"

That is entirely dependent on the OS. Here's a snippet from the "zpool" man page as written by Sun and distributed with FreeBSD:

EXAMPLES
Example 1 Creating a RAID-Z Storage Pool
Example 2 Creating a Mirrored Storage Pool
Example 3 Creating a ZFS Storage Pool by Using Slices
Example 4 Creating a ZFS Storage Pool by Using Files
Example 5 Adding a Mirror to a ZFS Storage Pool
Example 6 Listing Available ZFS Storage Pools
Example 7 Destroying a ZFS Storage Pool
Example 8 Exporting a ZFS Storage Pool
Example 9 Importing a ZFS Storage Pool
Example 10 Upgrading All ZFS Storage Pools to the Current Version
Example 11 Managing Hot Spares

Each section explains exactly what it does and gives example command lines. In short, while Linux is historically not well regarded for its man pages, that doesn't extend to the format as a whole.

Re:Finally...

[Hurricane78]

Man pages are the reference.
Every Software has (or should have) a reference documentation, and a guide.
The short are called "HOWTO"s in linux. And the guides are located at tldp.org.
When they mean RTFM, they usually mean that you should read the guide at this site, or something similar (like the KDE help for KDE), and if that isn't enough, read the reference.

If this does not solve it, you still can write a mail to the developers, or if this is not possible, ask in a forum or irc chat.

Re:Finally...

[Hurricane78]

I hate those documents, where someone rambles on for 10 pages, about things like "to move your mouse, put your hand on the mouse... you know, that thing on the side of your keyboard. you have to push it. then look at the flat surface in front of you. do you see that little arrow on it. no not that on the right border. that is the scroll bar..." and so on.

The point that apparently not a single writer of documentation (or software for that matter) gets, is that you can't write in one level of detail, for all users.
So you have to have a method to allow the user to increase and decrease the detail on the fly. Eg by allowing you to press "+" and "-" on a paragraph, and by internally using some kind of markup.
Same thing with software. Let beginners use the mouse, show context help to everything, without being asked for it, and so on. In that context help, also show a more advanced method (like keyboard shortcuts). Then allow the user to slowly grow into an advanced mode, and then into pro mode. But only if he uses the software so much that he will remember this stuff. Because usually, the user will only learn and remember, what he has to, to be able to get what he want like he wants it.

How about we define a new help system for Linux. With a standardized file format, and multiple front-ends (which should also grow with the user).
That format would be a composite document, of HTML, and another set of tags to define the level of detail.
The front-end (eg a browser plug-in) would then allow users to zoom in on things they have problems to understand, and quickly skip unimportant parts, while still knowing that there is more.
Also, already seen content should be marked as such, so that somebody who wants to know every detail, can use it to check things off.

Re:Finally...

[fm6]

I hate those documents, where someone rambles on for 10 pages, about things like "to move your mouse, put your hand on the mouse... you know, that thing on the side of your keyboard. you have to push it. then look at the flat surface in front of you. do you see that little arrow on it. no not that on the right border. that is the scroll bar..." and so on.

So do I. Unfortunately, the writing style you describe is all too common. I don't use it.

I mostly write for really for fairly technical users, so I can get away with assuming that my readers know how to use a mouse and interact with a GUI. But even if I couldn't, I have techniques for conveying the trivia to naive users without driving the advanced users insane.

It's not that hard to write newbie instructions in such a way that the advanced users can easily skip over it. You put in a separate part of the document and refer to it. (You usually want to to do this anyway; inserting the same text over and over in a technical document makes it hard to maintain.) If, for some reason, the trivia has to go inside a procedure, you write the procedure to that the main steps are in bold, then all the boring prose explaining those a given step follows in ordinary type. Advanced users just read the bold parts, but the other stuff is there if the newbies need it.

The point that apparently not a single writer of documentation (or software for that matter) gets, is that you can't write in one level of detail, for all users.

Absolutely. But note that different manuals have different target audiences. Figuring out the right level of detail for my target audience is an important part of my job. Also...

So you have to have a method to allow the user to increase and decrease the detail on the fly. Eg by allowing you to press "+" and "-" on a paragraph, and by internally using some kind of markup.

As a matter of fact, some help engines already have this feature. (Microsoft's does.) I've never seen it used effectively, though. If I wrote a lot of naive user docs, and if the deployment formats I have access to supported these feature, I'd probably give it a try. (We deploy PDFs and HTML with little or no active content, but both formats could support this feature.) In the end, though, I'd probably go back to the techniques I've just described.

Not to be a Luddite, but non-techwriters always overestimate the ability of technology to solve documentation problems. Technology certainly helps (if you can get the funding to deploy it and persuade your fellow writers to use it; this is an ongoing battle) but in the end there will always be a huge human element in good technical communication.

Re:Finally...

[myvirtualid]

the above is simply not correct. Putting aside issues of writing style, quality, completeness, etc., along with the abomination called info pages, manpages ARE documentation

You are putting aside things that cannot be put aside. Far too may Linux man pages are brief pointers to the info info, and far too often the info pages merely rehash the inadequate content of the man pages.

To qualify as documentation, a man page must include sufficient detail to act as a reference, not merely a command line summary.

UNIX man pages were documentation - and were how I learned UNIX, back in the day. Generally speaking, the Linux user is better off with web pages than with man pages - too many Linux man pages don't cut it.

Re:Finally...

[MasterOfMagic]

I'm talking about the manual pages, not the comments in the code. If a programmer waits to document the code that they are writing, then you are correct, they are bad programmers.

However, waiting to write a manpage, tutorial, or quickstart is something that can be done after the program works properly. Otherwise you've got a manpage, tutorial, or quickstart that will have to be constantly updated as you make changes to the program.

I've written Perl scripts that were well commented and became the backbone a small, end-user system. What was the last part of the deliverable that I worked on? A help guide - I wanted to make sure that my code worked and worked well. The code was always commented, but it was more important to me that it worked and maintenance programmers could work on it. The help guide was something I put together for the users, but it was the lowest priority.

Your comment reminds me of something I like to call "Whiny End User Syndrome". You are getting free software - software that you can use as you please, software that comes with the source code so you can study how it works, software that you can redistribute freely, and software you can modify and redistribute the changed version of. This software doesn't spring out of the ether, fully formed, tested, and documented for the end user - that takes time. Certainly I'm not saying that developers should skip writing documentation for the end user out of some malice for end users, but documentation (specifically end-user documentation) takes time away from coding and making your program more capable or bug-free.

In the end, either way, the programmer is screwed. If he spends a lot of time on documentation, there are a lot of users that cry, "Focus on features! We want more features and we don't care that it comes at the expense of the docs!" Then, when you spend a lot of time on features instead of documentation, these same users cry, "This is a great feature, but we need more documentation!" It's twice as bad when the code you've written is put out there because it's a clever hack and you never intend for end users to use it; you've put it out there because you want it to be available to the next poor soul that has to do the same thing you did.

So no, calling every developer that puts end-user documentation off until they're done making something stable a bad developer is painting with a mighty broad brush, and is unjustified.

For the rest of us...

[geekmux]

Something this basic requires a posting on the front of Slashdot? Great choice, kdawson!

I'll be sure to try and get my article about setting Openfire up on FreeBSD here soon..........

Ah, feel free to excuse yourself at anytime if you feel the rest of us are not worthy of your all-knowing power.

He was merely posting as a response to several queries he had received for the information, and since it's not quite as simple as apt-get to do this, along with the fact that FOSS collaboration tools are gaining popularity in this economy of ever-shrinking budgets, I find it rather relevant.

Re:For the rest of us...

[digitalchinky]

I don't like to add a "me too" but not just 5 days ago I added an "ejabberd" service. Found it to be fairly complicated and problematic. The documentation presumes one would be all knowing on the subject. Transports were complicated, modules assumed I would read the source code to make them work, the included readme files were a little bit circular and devoid of actual configuration information. And the init script that was provided completely failed to do anything at all. It took a few days messing about to get it somewhat usable, now this isn't a bad thing, I gained a good appreciation for instant messaging in general, but I run across this story just now, and here I am with openfire up and running, doing everything I wanted and much more, just 15 minutes later.

I too found this story rather relevant.

Re:For the rest of us...

[palegray.net]

You just put a big smile on my face. Your post is the reason I write stuff like this, in spite of those who seem all too willing to decry such efforts. Thanks :).

Re:Debian

[skyride]

Whats wrong with debian? Yes, CentOS and the likes are great but Debian is still a very good OS and its really just opinion of whether you prefer the APT or RPM system. Personally I use both. I have a dedicated server currently running a small number of Source-based Game servers and my home server is running CentOS 5.2. Debian Lenny is a great OS, its not really a complete re-write, but thats the point. Its a good, mature system which can only get better. And what operating system were you referring to? Windows? Sorry but most people aren't happy to buy Server 2008 for a small project like this. My guess is that more than half the people following this tutorial will be many who, like myself about a year ago, were keen on learning about linux.

User Gaming - xep-0196

[Midnight Thunder]

Since we are talking XMPP/Jabber, I would be interested to see the User Gaming specification implemented by XMPP messengers. This would make a nice open cross-platform alternative to XFire and the likes.

I noticed someone elsewhere suggested implementing this and then grafting a XFire compatibility layer around it, so that people could migrate to the open platform.

Is this something that would interest any /.ers?

Re:User Gaming - xep-0196

[skyride]

Greatly. Although I think steam has already made most of the features of xfire entirely obsolete. If it implemented something like a ventrilo style server based chat system it would be fantastic and an instant win.

Re:Google Talk Still experimental?

[Arkem Beta]

The article is referring to the connector that allows your server to talk to the Google Talk server to pass messages.
My guess is that it's unstable because Google Talk uses some non-standard options when talking XMPP.

Re:Debian

[Jacked]

Debian is popular for servers. Admittedly, I only admin a half-dozen boxes, nothing compared to what some of you guys work with. But, they work great.

Which others did you have in mind?

Re:Debian

[skyride]

My guess is he was referring to windows Server 03/08 or an RPM based linux, i.e. CentOS or RHEL.

"Collaboration"? "More than chat"?

[jnetsurfer]

TFA states, "[openfire is] an extremely capable collaboration system [...] able to do much more than simple chat". I use Openfire, and am not aware of it's collaboration system. Can anyone explain? SInce TFA doesn't seem to...

Re:"Collaboration"? "More than chat"?

[Anonymous Coward]

Jive software has 'collaboration' tools. They all suck, but they have them.

"Jive Software has always envisioned one world, in which everyoneâ"employees, customers, partners, prospectsâ"has a voice. We've strived to eliminate the top-down, outbound, one-way communication that dominates most enterprises and makes work harder than it needs to be."

And that is why they make closed source software!! ... wait ...

Re:"Collaboration"? "More than chat"?

[jnetsurfer]

LOL. Got it, thanks for clarification! FWIW I really like Openfire. Except for the memory footprint but that's due to its Java implementation...

Will they ever fix the audit log issue?

[mgkimsal2]

http://www.igniterealtime.org/issues/browse/JM-1212

This bit me last year, and it's apparently still not fixed. :(

Google Talk

[hax0r_this]

Try the Google Talk client (Google Talk is Jabber/XMPP), I don't know if you can connect to non-google servers, but with federation you should be able to talk to people on other servers anyway. Disclaimer: I don't actually have a BlackBerry, but I've heard the client is good.

Why read a how to, when you can watch it?

[Uteck]

Hak5 http://hak5.org/ just did a segment this week about this. Informative, entertaining, and not OS specific.

Re:Debian

[h4rm0ny]

What's wrong with Debian. Explain yourself.

Why not OpenJDK?

[Florian Weimer]

Why don't you install OpenJDK instead Sun's proprietary Java implementation? Is there a technical reason for this choice?

Re:Why not OpenJDK?

[lgbr]

Sadly, Openfire did have some issues when I used with icedtea6. The best example I can think of is the MSN transport. MSN simply wont connect because of the security algorithm it uses. This is caused by icedtea6 missing that algorithm which Sun's JDK has.

Re:Why not OpenJDK?

[Florian Weimer]

Sadly, Openfire did have some issues when I used with icedtea6. The best example I can think of is the MSN transport. MSN simply wont connect because of the security algorithm it uses. This is caused by icedtea6 missing that algorithm which Sun's JDK has.

Hmm, have you tried it with Debian's openjdk-6-jre package? I think the Sun security provider was opened quite some time ago.

Do you remember which algorithm causes the trouble? This is really a bit odd because even the Sun-derived JDKs (like those from IBM) differ a bit in their security providers.

Re:Why not OpenJDK?

[lgbr]

It actually seems to be a parameter issue. Here is the log output: http://pastebin.com/f6a134c54

This is gentoo's icedtea6:
IcedTea6 1.3.1 (Gentoo) Runtime Environment (build 1.6.0_0-b12) OpenJDK Client VM (build 10.0-b19, mixed mode)

Re:Why not OpenJDK?

[palegray.net]

As other posters have noted, I encountered issues with OpenJDK while I was writing the first draft of the tutorial. I'll probably go back and try it again, communicating the precise issues I ran across to the Openfire project. Hopefully they (or me, if I can find the time) can get the issues resolved, as I would have greatly preferred using a JVM in the "free" repositories.

Re:Debian

[cbiltcliffe]

Bad attitude?

BAD ATTITUDE?!?

I'll show you BAD ATTITUDE, you moron!!

Fricking cowards! They're all alike! Pop in just long enough to lob a comment grenade your way, then run off with their tail between their legs!

Shoot the bloody lot of them, I say!!

Now, onto more serious stuff....

BTW....Debian is a great server OS. Rock solid stable, some of the fastest security patching in the industry, and doesn't even attempt to get you to install a GUI on your server.

Re:Google Talk Still experimental?

[rabbit994]

GoogleTalk server uses XMPP standards to communicate. Only time I've seen Server to Server messages not function properly with GoogleTalk is when you do not have proper SRV records set for your domain.

Sounds good to me

[QuincyDurant]

As a potential end user for a small business, Open Fire sounds good. I had never heard (or indeed thought) of such an application. The tag "openfiresucks" concerns me, however.

I would not be the one to install the server nor would I welcome the need for much maintenance or support. Can anyone tell me the pros and cons?

Why isn't it in Debian already?

[Danny Rathjens]

I was surprised not to find it when I did an "aptitude search openfire" after seeing a few people mention it in response to the ask slashdot. (My jabberd2 setup works well enough for my company anyway - I was just curious.)
This howto says it is GPL, then says to install prerequisites such as sun java from "non-free" and then download a .deb file to install. Why wouldn't it simply be added to the debian repo so installation is as simple as "aptitude install openfire"? Any idea if the company behind it has some weird policies preventing this?

The openfire website itself does not inspire confidence. The link to the changelog is a 404 and the roadmap is dated over a year ago. But I can't find any reason why it isn't in debian already. (not even sid)

And shouldn't this howto be put on tldp.org? or are people more interested in making ad revenue these days than helping contribute back to the linux community?

Re:Why isn't it in Debian already?

[palegray.net]

And shouldn't this howto be put on tldp.org? or are people more interested in making ad revenue these days than helping contribute back to the linux community?

Wow, dude. Way to be appreciative of others' work. As a point of fact, I have every intention of contributing this documentation to TLDP. I guess it's a crime to post it to my own site first, for own community to enjoy. When you write your own documentation, you'll be more than free to determine where it goes and when. Until then, how about losing the attitude?

Hak5 covered Openfire on Debian this week

[doulos447]

For those more "visual" learners...

http://www.hak5.org/

Re:Debian

[palegray.net]

If you're interested in an apt-based distro, consider Ubuntu.

I don't even know where to begin on this one, so I'll just jump right in. Debian's community is far from broken. It's extremely active, as anyone who bothers to subscribe to the mailing lists knows. Your note about Ubuntu is beyond laughable; you are aware that Ubuntu is based on Debian, aren't you? Wow.

Advising your employer to use Debian stable is a great move. Shit doesn't randomly change and/or break, the base install has a small footprint (great for building appliances), and their security team is top-notch. I always find it funny when I update my Debian and Ubuntu boxes; patches always come down for Debian first, followed by Ubuntu a day or two later.

Short version of this post: If you're managing any kind of Linux servers for a living, you should be fired.

Re:Debian

[turbidostato]

"Debian takes years, literally years, between updates."

Yes. Debian Etch: 2007; Debian Lenny: 2009

But how about its competitors?
Red Hat 4: 2005; Red Hat 5: 2007
SUSE 9: 2003; SUSE 10: 2005; no news about corporate supported SUSE 11 (only OpenSUSE, but I admit I can be misinformed about this).
Windows XP: 2001; Windows Vista: 2006

"If you're interested in an apt-based distro, consider Ubuntu. It has corporate support"

So you think corporate supported Ubuntus take much less, uh?
Ubuntu 6.06 LTS: 2006; Ubuntu 8.04 LTS: 2008. Surprise! Quite the same than Debian.

"CentOS is obviously the gold standard."

Of course yes... it is Red Hat rebranded as you probably know, so you can see how it goes a bit above: two years from its last release; two years average between versions. Quite near to Debian's mark.

"Its community is broken [...] If you advise your employer to use Debian, you're giving them bad advice. An oddball distro with almost no market share or community support, with very little in the way of contractor support."

So, up to now, all you have is an irrelevant fact (since all other "competitors" take as much if not more between revisions) some unsupported affirmations and a proveable wrong fact (if you call IBM, HP and a host of minor partners like my employer "minor support" you sir, are utterly wrong).

"It's 2009, you can't just keep using what you started with, you have to stay current."

What do you think the 70+ Debian servers under my belt are exactly so uncurrent and compared to what, may I ask?

"If you're not willing to do that, get out of the way."

If all you can do is rude unstated and wrong assertions, you will find yourself out of the way wondering what happened... but of course I know that won't stop you, Mr. A. Coward.

Easiest solution

[Mr. Jax]

The easiest solution for setting up your own jabber server is to sign up for Google Apps with your domain (http://www.google.com/a/) activate talk and then you can connect with any Jabber client to the google servers, eg. with Kopete: http://www.google.com/support/talk/bin/answer.py?hl=en&answer=57557

You even have the option to limit communications within your own domain or to allow users to chat with other users outside of your domain.