Slashdot Mirror
Anatomy of a Successful Enterprise Linux Distro?
phenix asks: "With the new release of Novell Linux Desktop, and the upcoming release of Sun JDS3, I am curious to hear how these two suites, and their underlying enterprise infrastructures (JES and OES) compare. Specifically, I am interested in their ease of management/deployment in these areas: directory services, productivity (office) applications, centralized application serving, centralized document storage, groupware, and remote application installation. All of these, of course, without the use of Windows products like Exchange and Windows technologies like Active Directory. Is there a better alternative?"
If you wish to succed with implementing linux on desktops, inopperabillity Should not your main concern.
Office jocks are lazy technophic grunts. All they require is the fact that they can use it. A *stable* And most importantly usable GUI environment should be available. And sorry for the the blasphemy, but you might want to implement a GUI similar to windows. It will make the desktop feel more "natural" for it's users.
perpetually dwelling in the -1 pits
This is just a veiled "which distro is best" post (most of what poster asks is at application level, not distro level), this entire discussion will be a flamefest.
At risk of wandering offtopic, the parent post is a microcosm of why people don't try OSS. A user asks for help/opinion, and it gets thrown back at him/her. Almost as un-helpful as "RTFM".
"Ask not what your country can do for you." --John F. Kennedy
A successful enterprise Linux distribution?
Mac OS X + X11 + Apple Developer Tools (Xcode) + Fink
In all seriousness, we have found that a desktop or laptop with Mac OS X, with X11, all of the compilers and development tools, and a ports/package manager like Fink or DarwinPorts, which still allows running normal productivity software like Microsoft Office, mainstream media players, Adobe products, etc., has been the most productive platform of all.
The problem with Sun and Novell is that they are both approach the problem: how to be Microsoft. As such they are trying to be all things to all people (their own OS, own directory services, own productivity, etc.). I support the effort, but there are two many years and too much functionality of built up Microsfot competitive status to comprehensively replace in one package.
A more feasible / successful approach the "assimilation" that is being led by Xandros. Let the user keep his productivity suite (Crossover), keep his Active Directory (Xandros authenticates against it), keep their NTFS, etc. Above all, get the home and corporate user on the right OS (Debian in Xandros' case) and migrate the other functionality in a best of breed fashion in the future (when it is easier).
In some of the cases, Xandros did build out functionality that Linux normally lacks: e.g. remote application installation. In this case though, they also built Windows hooks so the same manager can control both Linux and Windows boxes: clever.
Hunger is the best sauce.
It comes with Samba 3.0 for SMB/CIFS, Active Directory authentication and a Microsoft Exchange connector.
Citrix and Acrobat Reader, OpenOffice2.0 etc
Hmm.. what else... NUMA support for multi CPU (also a lot of multicore enhancements)...LVM2 for easy disk addition, removal....
RHEL4
Unfortunately maybe not... I always use Linux server side, but have a Win2003 VMWare session running active directory. Why? It just works. It's all integrated... you add a user, set their password and kerberos and ldap work across the entire network. I don't have to *know* that it's ldap and kerberos under the hood...nor should I have to.
Sad to say there's nothing close to that for Linux yet (until we get a native version of NDS).
Farting around with ldif files and obscure samba configurations just doesn't come close.
What an odd question. NDS has existed almost 10 years, providing centralized appliation serving via NDS integrated applications. Look at Pegasus Mail (I assume this is what you mean). Install the app on the server, and the programs INI files are stored in the user's home directories. Users can move from PC to PC without migrating anything that's PC-specific (such as the registry). Hell, if there's any reason to get MS Source code, it would be to get the source to Outlook and rip out the registry crap.
Zenworks takes care of the rest of the desktop 'distribution', like installing and upgrading pc-centric software.
I would guess you didn't know Novell's Border manager could be thought of as IPChains based on NDS login.
It sounds like you really don't know much of what is out there, and you need to read some whitepapers at Novell.com. Or goto some tradeshows. Get exposed.
"I can't give you a brain, so I'll give you a diploma" - The Great Oz (blatently stolen sig)
I think that Linux Distributors need to help get distrobutions configurations optimized.
/etc/passwd and /etc/shadow Just don't cut it anymmore.
In Linux, often, the needs of the Enterprise have to translate into the needs of the Home User. I know you likely think that the Home User doesn't need OpenLDAP, when in reality, with the amount of information they have to manage using computers, they absolutely need OpenLDAP, MySQL, Samba, and other things.
Alot of Linux "Bugs" are fixable out of the box configuration issues. I have a friend of 8 years who much to my emotional devistation, is moving from Linux to Windows XP. The major issue he had? There was always something wrong in how something was configured.
The permissions not being set right on the CD Burner, Gaim not being absle to direct connect from behind a NAT, even a well configured Shorerwall NAT.
Linux can be configure such that it does "Just Work(tm)." The issue is the distributors, even Mandrake do a hard time gauging what the real needs of the Enterprise and Home Users are.
This isn't a "Linux Software is inferiror" issue its a "Why did you set the CD Burner to 600 when it should be 660" issue. These configuration issues cause Linux to fail. Giving people the impression Linux Software "Doesn't work" Like my friend.
Linux Distributors Underestimate the needs of Home Users and Distributors of this day and age with half-hearted configurations and sometimes downn right "Wrong" information. They substitute Universal comprehensive Linux Applications like Linuxconf for Proprietary ones like thee Mandrake Control Center.
The Distributors need to start creating more dynamic and sophisticated DEFAULT CONFIGURATIONS to meet the growing dynamic needs of today's home and Enterprise Users.
Free Software is not about being whiny and helpless.
Even Apple is not about that despite the fact that they've been the vangaurd of user friendly computing since foundation.
A Pirate and a Puritan look the same on a balance sheet.
This is my wish for a linux distro. I work in a Windows only company where linux is slowly creeping in (embedded, specialised application servers, a small beowulf). I have ZERO chance of getting admin rights on the network or influencing the IT dept. towards Linux. What I need is a distro which not only coexists with windows, but automatically sucks its settings, copies email settings, home drive and printers. Automatically install the correct autentication module so people can log on with their active directory password. Read the windows drive letters and mount them. Note I write this from a linux box. I am trying, but it could be easier.
10 ?"Hello World" life was simple then
Or maybe the poor guy is aware that sometimes issues crop up during a full deployment that don't show up in a test environment, and just wants to see what issues or concerns others have come across in their deployments.
:). But it could happen.
Or perhaps he just wants to supplement his testing with other opinions of the products which might bring to light issues that he hadn't thought of.
I'm a pretty pessimistic guy but it amazes even me the amount of negative feedback that is generated whenever someone asks a question.
Maybe he is just being lazy and not wanting to try it on his own but you don't know that. It's pretty rude to slam him on the basis of something you don't know anything about.
Not to mention the fact that by posting it in a public forum, he's bringing an important topic into public discussion which could produce a productive idea or two on the topic. Sure I know it's a slim chance, this is Slashdot after all
ender-
Nothing to see here
The other extremely important factor is the "Warm and fuzzy feeling(tm)" in your manager. It boils down to covering his ass. Having a really big name behind your distro really helps. Novell/Suse have a winning hand here. We all know that in the real world vendors won't cover jack-shit but it's all part of the big lie. Various bogus partner programs can also help here but not as much.
The more insecure your maneger is the more important these factors are.
But having your key applications certified on your distro of choice is essential unless you've got source for them and can roll them yourself.
TCAP-Abort
Name 3 things wrong with "their network security".
"Most of these features you wish to compare are not part of a Linux operating system. Most are applications that are installed on top of a Linux operating system.
So, among the Linux distributions, all of these features are roughly eqivalent, providing that you are using the same software to meet the need for the particular feature."
VERY untrue! You will find that, for example, LDAP support (and the ability to talk to AD if needed) varries widely from distribution to distribution, even though they all have roughly the same tools for talking to LDAP. The key questions are: how easy is it to set up; how well are applications, tools and libraries configured to integrate any given feature you need; are conflicting services installed by default (and/or REQUIRED);etc.
You might not think this way because you worry about <1000 end-user machines, but let me tell you, when you need to install 5,000 end-user desktops, you're not thinking, "eh, it's ok... I'll just install anything and configure to taste," you're looking for something that gets you as close to the finish-line as possible so that you can worry about the truly hard problems. Sure, you're just going to dupe hard-drives, but that doesn't get you the perfect economies of scale you might have expected. You're going to worry about things like, "oh look, this LDAP server falls over when 1000 clients ask it a question at once," and other scaling issues. You don't want to have to start at ground-zero, "why doesn't LDAP work with applications X and Y and works half the time with Z."
The above is just an example, but I think it illustrates the point.
The natural urge is to bar membership to this community to perpetuate the elitism, greatly harming new user adoption.
Sometimes. But untrue at least two-fold.
First, much of the FOSS community is genuinely interested in helping people to use FOSS, regardless of their abilities. The majority, in all likelihood. It's only a small fraction of FOSS users with some intelligence and more than a few personal insecurities that belittle the attempts of others to learn FOSS software.
Second, there's always the cachet associated with doing what's new and what's cool and what's not common. There are some who will actively seek to learn FOSS for the same reasons that they have the latest PDA gadget. The community grows in size somewhat because of these new members who usually delight in "explaining the mysteries of the universe" to co-workers, friends, the girl cashiering at the grocery store, etc.
The biggest problem is that enterprise wide deployments are the last frontier for Linux (and even that is rapidly being colonized). In the old days people complained about lack of desktop environments, lack of hardware support, lack of embedded processor support, etc. a whole lot of features which have faded in importance.
That's why the last battle will be the one for a distributed global authentication and authorization system, something which takes a while for the FOSS to come up with the standards and the implementation of an agreeable system. The established players (MS, Sun, etc) will just distribute their solution as part of their distribution.
"Provided by the management for your protection."
Utter nonsense.
One of the big causes of switching to Linux is that the maintenance requirements are much lower than that of windows, a single linux admin, ie Me can easily look after and maintain 70 servers and 200 desktops without too much grief and still have time left over to help out the windows admins and do some development work on the side and maintain the documentation.
Now we roll our own, and all of a sudden I have a nightmare on my hands, _thousands_ of packages have to be maintained by hand, testing requirements go through the roof, and when I have the nervous breakdown and get carted off to the loony bin, the poor sod who is my replacement is completely lost, because the documentation is months out of date, the servers will be having the seven shades kicked out of them by the l33t script kiddies and the users will be deserting KDE in their droves back to our win98 site license and the bosses will be looking for someone to (rightly) blame.
I suggest you actually carry out a couple of large scale conversion before making idiotic comments like this that only server to show how ignorant you are of Enterprise requirement.
Oh wait, sorry I forgot this was slashdot.
Look at Apple *AFTER* 10.4.
I'm running it for a cluster. LDAP was easy to set up, Kerberos a nightmare (the "you haven't connected to a domain" after you really have was a nice touch), and NFS is pretty doggy compared with AFP, while AFP makes SSH users go through gyrations to get their home directories.
Now, for what I do with it (crunch numbers and provide an office worth of desktop connectivity), it's a good system. If I were running a small web-farm, or moderate sized office, then, yes, i'd recommend it in a minute. It was leaps and bounds ahead of setting up 2003Server for the same purposes. However, in any kind of large enterprise, it needs some time to mature, or you're going to spend as much time becoming proficient in the OS-X/Next way of doing things as you would becoming a RHCE/HP-UX/AIX-jock.
Quite seriously, the question should be, do you want Linux for particular support reasons (pricing, software available, Linux-jocks in good supply, cheap hardware), or should you be looking at a more mature Unix, with more aggressive vendor-support? I know people who work at Enterprise-level sites, and they swear by HP-UX/Solaris/AIX.
the more accurate the calculations became, the more the concepts tended to vanish into thin air. R. S. Mulliken