Slashdot Mirror

I'm not much of a network admin, but Samba is a well-supported suite for including Linux boxes on AD domains, and I think has also gotten to the point where Linux boxes can act as domain controllers. So you can in fact start pulling in a few Linux boxen into your network without disrupting everything.

In addition, you may find that Linux gives you other tools that allow you to handle the stuff you currently use AD for.

Guy, literally every single one of your guesses is absolutely way out.

That is a wild exaggeration, for example his observation about copy-on-write process memory is absolutely correct and dates back to the dawn of Unix.

They're talknig more about classloaders and signed apps than ACLs and C++-style object models. Your SQL example is hilarious - no, they're not trying to count anything. They're creating shared class files originally comprised of greater than one class file, recognising where the byte blocks are identical. Think more VM or storage deduplication, not counting things.

There is more than one way to skin a cat. I would suggest running the byte code through rzip, which will deduplicate it even more effectively.

rsync
Samba
and of course this little gem

I expect at least one follow-up WOOOSH! post but I derive pleasure from putting smartarses in their place.

rsync
Samba
and of course this little gem

I expect at least one follow-up WOOOSH! post but I derive pleasure from putting smartarses in their place.

Self replies are bad, yada yada ...

The more I read the original post the more I think that slashdot has turned into a generic 'ask about your computer problems' forum than news for nerds.

This question appears to be from someone with no administration experience at all, nor does it appear that even the slightest effort has been put into finding a solution to the problem. Just for reference, telling us that you googled without actually doing it doesn't count as making an effort.

Its unacceptable that you don't know that Samba is your file server software and you're asking about how to configure it on slashdot. Yes, I know, I should be helpful and give you a bit of knowledge blah blah blah. You should also have not been so lazy as to not bother to even find out what software you are running before asking about how to configure it.

Whats more unacceptable is that douche bag let put this on the front page. I'm sure there are plenty of cluebies that send things to slashdot, but even timothy should know at this point that OS X uses samba, its not like it hasn't been discussed here multiple times before.

But finally, just so I'm not 'that guy' who's a totally useless asshole.

http://us4.samba.org/samba/docs/man/Samba-HOWTO-Collection/speed.html

That guide is for making samba faster, figure it out.

Microsoft did nothing "nice". They were dragged, kicking and screaming, into court and had their fingers slapped to the tune of over one billion US dollars by the EU for their misbehavior. And they attempted to poison the well by inserting patents into the published documents, patents incompatible with GPL software such as Samba. There are plenty of references to the court cases, but the interview with such developers of Samba as Jeremy Allison at http://www.groklaw.net/article.php?story=20070919214307459 are particularly enlightening.

The Samba site also has this note about the patent encumberment and GPL incompatibility Microsoft tried to slip in: http://us1.samba.org/samba/ms_license.html.

And if you think there's anything "nice" about their efforts, go read the documentation. It was apparently written by monkeys trying to produce Hamlet, and bears little if any resemblance to how the protocols actually work.

http://samba.org/samba/PFIF/

Samba and any other free software project (via the PFIF) has a royalty free license to most of the patents that are important for these protocols.

There are some patents that are excluded from this (see appendix 4 of the agreement for a list of the excluded patents), and we do indeed need to avoid infringement of those patents. That has not so far proved to be an insurmountable obstacle, although it is an inconvenience.

Cheers, Tridge

The next problem is the "make" mindset, which is built on timestamps. "make" doesn't check what changed; it checks was was "touched". If "make" decided what had changed based on hashes, rather than timestamps, many unnecessary recompiles would be avoided. Something could run "autoconf", produce exactly the same result as last time, and not trigger vast numbers of recompiles. .

Check out ccache. Basically you symlink 'gcc' to it. Before running GCC on a source file, it actually computes the hash of the source file and sees if it was compiled already. If it was it just fetches it from the cache. Why this facility isn't included as standard on every system is a mystery to me.

I had a similar experience many years ago - my rant was archived for posterity thanks to the local LUG mailing list archive. Full story is here: http://lists.samba.org/archive/wireless/2003-May/002533.html

WTF? How can you possibly justify your position?

Lets just a quick "Lets get the facts straight campaign":

A 2003 license is $429.99 US ex tax (Euro pricing, I am sure that the US is cheaper) and that includes 5 CALs. Datacentre runs well and truly above your $3,000 figure, try doubling it if you want Hyper-V.

A 2008 CAL is about $30, but it's not just that you are probably going to want, it's sharepoint and everything else. So really, you just haven't done any research.

Lets run with your understanding about using Linux to connect to Windows, it's wrong.

If you aren't using their software, why would you have to pay for a Client Access License? I am sure you could make a donation to the Samba Foundation, and I am sure that they would appreciate it. Aside from that though, why would the protocols need a license? They have publicly posted the protocols, they got forced to by the EU as part of their anti-trust investigation. This was part of their settlement. They have also posted the protocols for Exchange and a number of other protocols; they had to.

Really, this is the whole point of Jeremy Allison going tot he EU hearings and testifying and everything else, to MAKE Microsoft go through the interoperate with everyone else. Take a look here: http://www.samba.org/samba/PFIF/PFIF_history.html

Disclaimer: I am not an apologist, I am a Linux advocate but I still use a lot of MS products in my day to day business

If you follow RMS's lines, the mpeg group is just as evil because they have patents on their codecs.. wine should be completely banned, and hell, even openoffice implements a 'standard' made by microsoft in the name of compatibility (mono implements .net ecma standard also)

Samba would also have to go too, and the list goes on. All of these projects were made completely independently by enthusiasts that just wanted to make things for compatibility sake

Yeah except that's not RMS' "lines". He's not against interoperability or compatibility.

Wine doesn't implement any patented technology -- if it did, it would be a ridiculously easy target. All proprietary and patented codecs etc have to be acquired by the user. RMS' only problem with Wine is that it is largely used to run proprietary Windows software -- using free software to remain locked-in to non-free software. But the software itself is free and no danger to include in a linux distro.

Samba is similar. That's actually a project RMS loves, because it undoes proprietary lock in. They didn't implement anything patented. And when they had to in order to get all of the functionality, they did so under an agreement that was in accordance with the EU antitrust ruling.

If Microsoft's Open Specification Promise had been overseen by the EU, then maybe it wouldn't have a provision where the "irrevocable" agreement may or may not apply to future versions and that it doesn't apply to any "non-conforming" software.

If every potentially patent encumbered item were removed from a linux distro, there wouldn't be anything left, even ancient things like fat drivers still have some patent litigation around them.

Yeah, because there aren't any non-proprietary non-encumbered file systems out there, FAT is it. *giant eyeroll* If you take "potentially" to mean "realistically" rather than merely "hypothetically", then there would be plenty left. Like... just about everything in a modern distro. There were "realistic" patent problems with writing .gifs. So libgif only supported decoding them. Zomg, that leaves... only all the other image formats! You're right, that's, like, nothing!

This almost non-sensical fear of mono seems strange to me.. I consider it like wine

Well there's your problem. It's not like wine. That's why it doesn't make sense to you.

The ONLY way you could consider it non-free software is if you count anything that could possibly infringe patents...

Or if you CORRECTLY view it as something that DOES infringe patents with 100% certainty, and is only as free as Microsoft's on-going good will towards free software. Then you'll realize why everyone in the know is concerned about this project, and not Wine or Samba.

If you follow RMS's lines, the mpeg group is just as evil because they have patents on their codecs.. wine should be completely banned, and hell, even openoffice implements a 'standard' made by microsoft in the name of compatibility (mono implements .net ecma standard also)

Samba would also have to go too, and the list goes on. All of these projects were made completely independently by enthusiasts that just wanted to make things for compatibility sake

Yeah except that's not RMS' "lines". He's not against interoperability or compatibility.

Wine doesn't implement any patented technology -- if it did, it would be a ridiculously easy target. All proprietary and patented codecs etc have to be acquired by the user. RMS' only problem with Wine is that it is largely used to run proprietary Windows software -- using free software to remain locked-in to non-free software. But the software itself is free and no danger to include in a linux distro.

Samba is similar. That's actually a project RMS loves, because it undoes proprietary lock in. They didn't implement anything patented. And when they had to in order to get all of the functionality, they did so under an agreement that was in accordance with the EU antitrust ruling.

If Microsoft's Open Specification Promise had been overseen by the EU, then maybe it wouldn't have a provision where the "irrevocable" agreement may or may not apply to future versions and that it doesn't apply to any "non-conforming" software.

If every potentially patent encumbered item were removed from a linux distro, there wouldn't be anything left, even ancient things like fat drivers still have some patent litigation around them.

Yeah, because there aren't any non-proprietary non-encumbered file systems out there, FAT is it. *giant eyeroll* If you take "potentially" to mean "realistically" rather than merely "hypothetically", then there would be plenty left. Like... just about everything in a modern distro. There were "realistic" patent problems with writing .gifs. So libgif only supported decoding them. Zomg, that leaves... only all the other image formats! You're right, that's, like, nothing!

This almost non-sensical fear of mono seems strange to me.. I consider it like wine

Well there's your problem. It's not like wine. That's why it doesn't make sense to you.

The ONLY way you could consider it non-free software is if you count anything that could possibly infringe patents...

Or if you CORRECTLY view it as something that DOES infringe patents with 100% certainty, and is only as free as Microsoft's on-going good will towards free software. Then you'll realize why everyone in the know is concerned about this project, and not Wine or Samba.

The Microsoft Community Promise is not good enough. See this legal analysis for details:

http://www.softwarefreedom.org/resources/2008/osp-gpl.html

Microsoft lawyers are good enough to produce a better document than this, they just chose not to. See this document:

http://www.samba.org/samba/PFIF/PFIF_agreement.html

for a better agreement and an analysis on why all the terms in it are needed (especially the "Patents" section).

Jeremy.

they don't like introducing "new" things

A slight correction, they like to introduce new things when it suits them. Why the rewrite of SMB into SMB2? Well, it has some technological advantages you would expect but according to Wikipedia:

SMB 2 has two big benefits to Microsoft. The first is clear intellectual property ownership. SMB 1 was originally designed by IBM and was shipped on a wide variety of non-Windows operating systems such as SCO Xenix, OS/2 and DEC VMS (Pathworks). It was partially standardised by X/Open and also had draft standards for IETF which lapsed. (See http://ubiqx.org/cifs/Intro.html for historical detail).

The second benefit is a clean break. Microsoft's SMB1 code has to work with a huge variety of SMB clients and servers. A large number of items in the protocol are optional (such as short and long filenames), there are many infolevels for commands (selecting what structure is returned to a particular request), Unicode was a later addition etc. With SMB2 there is significantly reduced compatibility testing (currently only other Windows Vista clients and servers). Additionally the code is a lot less complex since there is far less variability (e.g. there is no need to worry about having Unicode and non-Unicode code paths as SMB2 requires Unicode support).

So you can see they like to introduce new things when it means they have clear intellectual property ownership rights over it and also a lot less work for them. They also don't have to be backwards compatible with their own products.

While SAMBA 4.0 has experimental support for SMB2 interfacing, I'm guessing the "clear intellectual property" could spell trouble moving forward for Tridgell and the SAMBA team.

Huh... when did that happen? I've never had it work. I just checked in Snowy and it still doesn't work there. Which is sad since they use samba for the server. You'd think they'd at least match its client capabilities.

Of course, all CIFS is proprietary. The Samba team just does an awesome job reverse engineering it. The Open Group SMB standard is now basically a fork thanks to MS's policy.

OK, looks like Samba added DFS about two years ago.

All I can think of is lack of DFS support. But I think DFS is proprietary.

The Linux CIFS Client supports DFS, and Samba can serve DFS.

I haven't used ClearCase but I am wondering to what extent the features you mention can be provided using free revision control systems. Extremely lightweight branching is a standard feature of all modern systems; making a new branch for each feature or bugfix or experiment is pretty much standard practice with git. 'View spec' sounds cool. It could probably be implemented on top of git with a little bit of scripting (create a new branch, merge in the versions specified, then delete the branch when no longer needed). A make variant that reuses already-built objects sounds like ccache.

you forgot to mention that it also does a lot of things worse

Which you then fail to list?

...but requiring hardware produced in 2010s. switch the aero off and the remains will not even equal the clean nice look of windows 2000

Utter b.s., it runs perfectly well on a two year old laptop.

you probably meant most amateur computer users.

Yeah, you know, those people who buy and use a product and make it profitable and the de facto standard? Yeah them. Also me.

being hdmi is irrelevant. resetting display settings is not that annoying. moreover, most good cards come with utilities that overcome this xp bug and you switch the monitor and resolution on the fly from cards' utility

No, resetting the display settings IS annoying and, furthermore, bullshit. But that's what happens when you are using an ancient OS to run hardware that didn't even exist when it was made.

LOL. on my old 1GHz/512ram/pata hdd i have 22 seconds from ntldr to busy cursor gone. windows 7 doesn't even install on that

Funny you should mention installing, Win 7 installed in maybe half the time that XP takes for me. I don't know what you're doing at bootup but mine takes about 20-25 seconds in Win 7.

so does media player classic home cinema. even better.

I'm well aware of mpc homecinema, I use it as my primary video player. However, the media centre in Win 7 has a nice media library built into it which is designed to look good and work well on a plasma, and to be used with a remote if you have one.

then you're probably not using windows explorer at all. and right clicking on taskbar items to bring up the applications' system menu, neither. and the start menu shutdown-confirmationless item, also

No, I am not using IE, I didn't realise I had to use it if I wanted to use Windows.

does this (1Gb ram, 16 Gb hdd) look modest to you ?

Yes. Yes it does. It's 2009 for Christ's sake. You can build a system with 4 gigs of ram and a 500 gig hdd for a few hundred bucks.

... at making impression to home users. in enterprise, this is just a second vista. joining a samba/nt4 domain is a pain in the ass or impossible.

Amazingly, most of the people in these enterprises you speak of are the same 'casual users' you dismissed earlier.

Many 'amateur' astronomers are not amateur at all. And when the world's largest telescopes have full schedules targeting deep space mysteries and other weird objects, the 'common' celestial targets are ignored. Amateur astronomers don't ignore such things, they spend a lot of time and effort to observe them.

For example, the planet Jupiter.

Here's a discovery made by another amateur, Anthony Wesley. An impact mark on Jupiter, similar to Shoemaker-Levy which occurred back in 1994. And here's the link.

I didn't see anything about it in /. Why?

i'm see this photo http://jupiter.samba.org/jupiter-impact.html .I not have old photo to compare this photo

Here are the latest images at a more, um, robust location. Note the confirmation from JPL/NASA.

http://www.samba.org/rsync/

Or something like it in Windows.
It will check the delivered files, retry chunks if needed.
Much better than just FTP

I think the original poster intended to ask about a client application protocol. Like their reference to FTP. If my assumption is correct, then the poster from the article might be interested in something like rsync, only for Windows instead of Unix. In which that case would be cwrsync - Rsync for Windows.

I invested more time in making this post than researching file transfer solutions for this article. More research by the interested party would be wise.

Fortunately, at least in an AD environment, you can pull this off using GPO-based software restrictions. Unfortunately, Samba 3 doesn't support GPOs, and Samba 4 isn't quite done yet. That said, if it can be done via GPO, it can probably be done via the registry if you can figure out what keys they're munging, and once you figure that out, you can write startup batch scripts that can make the necessary changes to the registry for you.

I did some more digging. http://lists.samba.org/archive/samba-technical/2009-February/063187.html Amusing enough, the poster identifying the registry keys and the bugs in Samba works for Microsoft. So yeah MS did help fix the problem.

This email also identifies the registry fixes which compromise security in the name of interoperability. The fixes (disable secure signing, dropping 128-bit req, and allowing ntlm and lanman auth) qualify as a Cat-I vulnerability findings if you happen to be subject to network vulnerability scanning.

In other words, you're lazy because writing such an app takes effort and requires making decisions that you are too lazy to make. Fortunately, there are people who see the value in making that effort. Take Samba as an example. There are several GUI configuration tools available. All of these output a text file which samba uses as the config. You can use a GUI if you want, edit the text directly, or script it until your heart's content.

[...] I hope you'll accept the first batch of reasons as proof enough that your position is indefensible.

No one will buy that "indefensible" excuse. Other operating systems and their application programs do it, handle it just fine, and have done so for decades. The expectations of current and potential users have already been set and they demand ease of use. The majority of people who use computers are not going to open a text editor and start editing xorg.conf to add a second monitor or start hacking on smb.conf to share a folder. That's not an acceptable solution when other operating systems do a better job of catering to their needs.

If Linux advocates want to attract more users to their platform they will need to cater to the needs of those users. Us nerds are a small percentage of the computer using population and many of us are already using Linux. The growth in Linux users will have to come from the pool of regular users where using GUI tools is preferred. Linux won't attract those users unless it can prove to be at least as good a value or better to the end user. Application developers can work to achieve minimum parity with other programs in terms of ease of use or stick their heads in the sand.

Why are so many distributions still maintaining their own compile farms? Why not just fire up a pool of servers on EC2 and use http://distcc.samba.org/ to build all the RMS/YUM's/Packages/Whatever? Why not just fire up a bunch of extra web servers in the cloud when you push out a new release of your distribution? It is probably way cheaper than getting donated hardware and hosting.

Hell if all you haters were smart, you'd be pressuring the FSF to have its own "cloud" that GPL users could tap into as a compiler/testing farm. Your open source project could just fire up a server that is running your testing image and use it to create binaries or run automated tests.

Actually no, I'm a busy admin and I don't have time to follow these instructions for getting Samba hooked up to Active Directory: http://wiki.samba.org/index.php/Samba_&_Active_Directory

In addition, those instructions don't help with the #1 issue of UID mapping when you have multiple Samba machines.

When you use these instructions, Winbind grabs a list of Windows accounts and maps them to UID/GID values as would appear in /etc/passwd. The mappings are permanent at that point. If you later add a new user or group to AD, then the user or group gets the next available UID/GID.

The problem is when you join linuxone to the domain, get everything working, then add some AD users and join linuxtwo to the domain. Now, the two machines almost certainly have different UID mappings for the new AD users. This means that you have decent interoperability from Unix to Windows, but any users authenticated from AD have crappy Unix to Unix interop.

In addition, there are PAM modules that allow you to control access based on group membership. These work 100% (either group name or GID) with Unix groups, but cannot use AD group names. So, you have to use the GID that gets mapped, and the mapping can be different on each Unix machine. So, if you use a Unix workstation for users, you can't roll out a standard image...it must be run-time configured.

Last, if for some reason you have to unjoin the Unix machine from the domain and then rejoin it, the mappings will change. So, all those files created in the home directory of joeuser with UID 50021 will be unreadable when he logs back in and joeuser has a UID of 50043.

Documentation isn't an issue; implementing the protocols is sure to be hard work, that even a large corporation with a huge development team would have trouble doing (Windows' own implementation didn't get developed over night...):

Microsoft must keep the documentation up to date with new products and provide error correction assistance to parties signing the agreement. Disputes will be resolved by the Trustee appointed by the Commission as part of the court decision.
...
The patent list provides us with a bounded set of work needed to ensure non-infringement of Samba and other Free Software projects that implement the protocols documented by Microsoft under this agreement. Any patents outside this list cannot be asserted by Microsoft against any implementation developed using the supplied documentation.
...

Actually no, I'm a busy admin and I don't have time to follow these instructions for getting Samba hooked up to Active Directory: http://wiki.samba.org/index.php/Samba_&_Active_Directory

Then I have to install ACL support and headache that goes with that, hoping something doesn't scramble my file system. In most businesses, Windows Server is not terribly expensive and allows the admin to get more done in less time.

Note, there are distros that offer GUIs for getting this done but they generally cost $$$. Why spend $$$ and introduce multi OS environment into what used to be single OS environment resulting in additional headaches.

Yeah, I know. Thankfully a new installation is safely locked down so that you can only browse the Microsoft website. Imagine what might happen if you could browse the web freely. You might accidently end up here which everybody knows is a site full of trojans and malware.

I agree with your comments and want to add the following Win7 debilitation:

I stopped using the Windows 7 beta, even though for the most part it is less offensive than Vista, due to it's inability to join a samba nt4 style domain. See here for MS confirming this as an issue with utmost importance: http://lists.samba.org/archive/samba-technical/2009-January/062827.html

If this gets fixed, I would be happy to move away from the steaming pile of Vista, to the merely warm pile of 7, but given MS' track record with compatibility, I'm not holding my breath.

Jim Pinkerton asked me to post this message from Microsoft:

Sincere apologies, but wanted to confirm that there is an issue with NT Domain support in Windows 7. We're trying to expeditiously track down exactly the issue(s), but the short of it is I wouldn't spend time trying to get this functionality to work. We'll get some sort of official notice out shortly (and I do mean shortly).

Jim Pinkerton
Microsoft

There still hasn't been a fix.

My guess: Microsoft made the process of joining a domain more LDAP based and reliant on Active Directory. Still, I just hope this "bug" is fixed.

2 days ago I installed samba 4 in a virtual environment. I donwloaded a debian lenny cd, installed a standard server plus some development tools (build-essential) and followed the instructions in the samba wiki: http://wiki.samba.org/index.php/Samba4/HOWTO. In half an hour I had a kerberos/ldap/rpc domain controller. Joining windows xp pro or win2k3 machines to this domain was a no brainer. I installed the standard microsoft tools (adminpak.msi, resource kit, resource tools) and I can now manage this linux AD from a windows xp with ADUC(dsa.msc), the dstools, the group policy manager, ...

So basically, any of my microsoft colleagues can manage this linux AD installation. Heck, they would not know that this is a windows AD except for the fact that they cannot login the server with the remote desktop client :-); I guess we should call it AD server linux core edition :-)

Samba people: THANKS!!!! The salvation is close.

This is a space that I've observed for a long long time. I can assure you that if anyone ever gets even remotely close to a replacement for Outlook against an Exchange server (or Exchange against an Outlook client), Microsoft will change the APIs so fast your head will spin off and fly away.

MAPI, AD and such are PROPRIETARY protocols folks, and Microsoft knows they are the keys to the kingdom. That's why all the Exchange clients ever created work ok at the start, but before they can really get going they fall back several steps.

http://blogs.zdnet.com/microsoft/?p=1064

http://arstechnica.com/open-source/news/2007/12/antitrust-pact-payoff-samba-gets-protocols-from-microsoft.ars

http://news.cnet.com/8301-13580_3-9836784-39.html

http://news.samba.org/announcements/pfif/

You are quite a bit out of date with that thought.

O? http://wiki.samba.org/index.php/Franky

I did post this link in another reply but it went aWoL... hope it helps clarify the story...
the other comments that I gave in that reply was that initially Microsoft asked for an exaggerated amount of money for the agreement fee (even though it's not mentioned in the link, I recall the news from a year ago) and that Microsoft is still holding back parts of the protocols with the "excuse" that they were enhancements done after the court decision, so that in reality the Samba team have to steer clear from these while trying to implement the protocols defined in the agreement that they paid for...
what's still not clear though is whether the end users will have to pay for the right to use the protocols as implemented in Samba, or they are covered also in the agreement after the further discussions, or if the pledge applies to the end users as well ?

Considering that SMB is an IBM invention they'll have to go against IBM.

That may be, but they are a lot closer now. The most interesting article I've seen so far is in this thesis.

mark my words, it'll have bugs which will result in 1000's of "RTFM n00b" or "it's ms's protocol that sucks" responses.

Just as Slashdot is full of trolls and OT comments help forums often have people posting unhelpful comments. Just ignore them. Life is too short for arguing with idiots.

I find the Samba help forums are generally excellent if you take the time to ask a sensible question instead of just posting the first problem that comes up. Often the task of formulating a sensible question solves a problem without actually having to ask on the forums at all. I also generally find my query has already been answered in the forum and all I need to do is search.

The Samba documentation is an excellent resource and generally answers most of the questions you may have. Try starting with John Terpstra's Samba 3 by example which is a practical guide to implementing Samba 3. I don't know if John is working on a Samba 4 update to the book, but there is a WIKI, HowTO and a FAQ available. If you are risk averse you may not want to use Samba 4 in production just yet :)

mark my words, it'll have bugs which will result in 1000's of "RTFM n00b" or "it's ms's protocol that sucks" responses.

Just as Slashdot is full of trolls and OT comments help forums often have people posting unhelpful comments. Just ignore them. Life is too short for arguing with idiots.

I find the Samba help forums are generally excellent if you take the time to ask a sensible question instead of just posting the first problem that comes up. Often the task of formulating a sensible question solves a problem without actually having to ask on the forums at all. I also generally find my query has already been answered in the forum and all I need to do is search.

The Samba documentation is an excellent resource and generally answers most of the questions you may have. Try starting with John Terpstra's Samba 3 by example which is a practical guide to implementing Samba 3. I don't know if John is working on a Samba 4 update to the book, but there is a WIKI, HowTO and a FAQ available. If you are risk averse you may not want to use Samba 4 in production just yet :)

mark my words, it'll have bugs which will result in 1000's of "RTFM n00b" or "it's ms's protocol that sucks" responses.

Just as Slashdot is full of trolls and OT comments help forums often have people posting unhelpful comments. Just ignore them. Life is too short for arguing with idiots.

I find the Samba help forums are generally excellent if you take the time to ask a sensible question instead of just posting the first problem that comes up. Often the task of formulating a sensible question solves a problem without actually having to ask on the forums at all. I also generally find my query has already been answered in the forum and all I need to do is search.

The Samba documentation is an excellent resource and generally answers most of the questions you may have. Try starting with John Terpstra's Samba 3 by example which is a practical guide to implementing Samba 3. I don't know if John is working on a Samba 4 update to the book, but there is a WIKI, HowTO and a FAQ available. If you are risk averse you may not want to use Samba 4 in production just yet :)

mark my words, it'll have bugs which will result in 1000's of "RTFM n00b" or "it's ms's protocol that sucks" responses.

Just as Slashdot is full of trolls and OT comments help forums often have people posting unhelpful comments. Just ignore them. Life is too short for arguing with idiots.

I find the Samba help forums are generally excellent if you take the time to ask a sensible question instead of just posting the first problem that comes up. Often the task of formulating a sensible question solves a problem without actually having to ask on the forums at all. I also generally find my query has already been answered in the forum and all I need to do is search.

The Samba documentation is an excellent resource and generally answers most of the questions you may have. Try starting with John Terpstra's Samba 3 by example which is a practical guide to implementing Samba 3. I don't know if John is working on a Samba 4 update to the book, but there is a WIKI, HowTO and a FAQ available. If you are risk averse you may not want to use Samba 4 in production just yet :)

"I'd just like to echo what a few other posters have suggested: stick with AD for now and migrate to Samba4 when it matures .. Samba4 will be a great drop-in replacement for AD but it's still some way away from being properly production-ready"

'In short, you can join a WinNT, Win2000, WinXP or Win2003 member server to a Samba4 domain, and it will behave much as it does in AD, including Kerberos domain logins where applicable'

Samba is freely available, unlike other SMB/CIFS implementations, and allows for interoperability between Linux/Unix servers and Windows-based clients.

"I'd just like to echo what a few other posters have suggested: stick with AD for now and migrate to Samba4 when it matures .. Samba4 will be a great drop-in replacement for AD but it's still some way away from being properly production-ready"

'In short, you can join a WinNT, Win2000, WinXP or Win2003 member server to a Samba4 domain, and it will behave much as it does in AD, including Kerberos domain logins where applicable'

Samba is freely available, unlike other SMB/CIFS implementations, and allows for interoperability between Linux/Unix servers and Windows-based clients.

Samba isn't a directory service, it's a Linux-based implementation of CIFS/SMB, and as such, is hardly "drop-in" replacement for AD. Why you got modded up for asking a question that reveals such a fundamental lack of knowledge is beyond me. But, this *is* Slashdot in the 21st century, so I suppose I shouldn't even bother asking.

True.. But you know, Samba 4 is actually supposed to include an ldap backend and will be quite near a drop-in replacement for AD.
It will still possible to use, for example OpenLDAP as the backend if one would like to.

Crap, that text has been in there since Samba3 was released.
I had hoped this would speed up since MS realeased that pile of documentation.

Apparently, all AD functionality will be put in Samba4, which is still Alpha:
http://wiki.samba.org/index.php/Samba4/FAQ

Samba could only be a DC on an old Windows NT style domain, not a Windows 2000/2003 style Active Directory domain.

No matter how you slice it, Samba is not a directory service.

See here:

Samba ADS Domain Control Samba-3 is not, and cannot act as, an Active Directory server. It cannot truly function as an Active Directory PDC. The protocols for some of the functionality of Active Directory domain controllers has been partially implemented on an experimental only basis. Please do not expect Samba-3 to support these protocols. Do not depend on any such functionality either now or in the future. The Samba Team may remove these experimental features or may change their behavior. This is mentioned for the benefit of those who have discovered secret capabilities in Samba-3 and who have asked when this functionality will be completed. The answer is maybe someday or maybe never!
To be sure, Samba-3 is designed to provide most of the functionality that Microsoft Windows NT4-style domain controllers have. Samba-3 does not have all the capabilities of Windows NT4, but it does have a number of features that Windows NT4 domain controllers do not have. In short, Samba-3 is not NT4 and it is not Windows Server 200x: it is not an Active Directory server. We hope this is plain and simple enough for all to understand.

Linux on the server side is a serious threat to Unix and Windows. First, they replaced Unix because they could do the job for a fraction of the cost except in cases like Big Iron where performance and reliability were the top priorities. Second, they opened up the minds of CEOs. Except for Exchange and AD servers, Linux is a serious competitor to Windows on most server applications. With the Samba team finally getting Windows Workgroup protocol information they always wanted, that AD advantage may slip.

Most people also haven't looked at what you can do with rsync and its backup option to preserve old versions of files. See http://rsync.samba.org/examples.html for examples.

Combined with connections to an off-site backup or then backing up the rsync destination, you get easy per-machine backups with versioning too.