Should OS X Server be licensed to other companies
I work with bunch of Windows Bigots, their biggest complaint about OS X server is that it only runs on Macintosh. What if Apple licensed OS X server to companies like IBM, Dell, HP/Compaq, sell it as alternative to LINUX, Windows.
If you had a choice of Dell PowerEdge with
1. RedHat Linux
2. Windows 2000
3. OS X Server
which one would you pick?
How many of you feel the same way?
If you had a choice of Dell PowerEdge with
1. RedHat Linux
2. Windows 2000
3. OS X Server
which one would you pick?
How many of you feel the same way?
Comments
But if that happens I don't think that Mac OS X Server will cost $999 anymore.
Originally posted by I.P.Freely
If you had a choice of Dell PowerEdge with
1. RedHat Linux
2. Windows 2000
3. OS X Server
which one would you pick?
Er, the first one?
Seriously, OS X is a consumer OS.
Originally posted by 123
Seriously, OS X is a consumer OS.
Seriously, tell me why you think that.
Originally posted by 123
Er, the first one?
Seriously, OS X is a consumer OS.
Ummm, have you seen OS X server? Just because it's easy to use doesn't mean it's only for consumers.
That, and since OS X Server is OS X with a bunch of extra applications, this lets the OS X on Intel cat out of the bag, which Apple might not be at all interested in doing.
Originally posted by JLL
Seriously, tell me why you think that.
1) Performance (mach overhead still there even though they put the whole thing into kernel space)
2) Reliability (this is from personal experience, I do something, suddently the machine becomes completely irresponsive for a minute or so. Way too many kernel panics, ever mounted a samba file system? Very instable)
3) The whole thing seems to be unfinished (Read some manpages, half of it contains stuff that's not there because Apple has chosen to not implement it. for example ipfw, it's supposed to provide traffic shaping mechanisms, in 10.1, all those commands were listed in the "usage" but they didn't work. In 10.2, instead of adding the functionality to the kernel (it's a kernel issue), they just removed the commands from the usage, but they are still there in the man page (I'm talking about pipe and queue). Or another example:
man mount
.
SEE ALSO
mount_devfs(8)
% man -S8 mount_devfs
No manual entry for mount_devfs
Really, the whole BSD stuff is just one big hack. Some things work, a lot don't.
And here's my favourite:
% cat /sbin/mount_ftp
#!/usr/bin/perl -w
#print "this is a test:" . @ARGV . "\
";
#for (my $i = 0; $i < @ARGV; $i++) {
# print $ARGV[$i] . "\
";
#}
So, the guy who has written this script didn't know argument lists and had to test them?
Then it goes on:
$items = 0;
for my $arg (@ARGV) {
$items ++;
}
WTF? He could have used a scalar assignment or scalar(@ARGV) or $#ARGV +1 or at least a simple for (@ARGV) or foreach loop.
I mean, if guys who write this also write the kernel, how can this be a good system?
Do this experiment:
mv /etc/hostconfig /etc/_hostconfig
reboot
Have a nice day, the removal of a simple!!! config file prevents the system from booting.
(when you see the bash prompt, remount the root file system as read/write:
mount -u -o rw /
and then put the file back to its old place.)
Why do I even know this? Because a few days ago, something (I still don't know what exactly, but I suppose it was the sharing control pannel) moved the file to hostconfig.old.
It's those little things that often make me wonder. I really believe that darwin is not mature enough to be seriously considered a server OS. And even without those issues, I can't see a single reason why I should choose it over linux.
Finally:
4) The user base is too small. If you have an issue, nobody can help youl.
(Though i prefer OpenBSD
#2 Linux .....good for databases (eg Oracle
my 2 c
Originally posted by 123
1) Performance (mach overhead still there even though they put the whole thing into kernel space)
2) Reliability (this is from personal experience, I do something, suddently the machine becomes completely irresponsive for a minute or so. Way too many kernel panics, ever mounted a samba file system? Very instable)
3) The whole thing seems to be unfinished (Read some manpages, half of it contains stuff that's not there because Apple has chosen to not implement it. for example ipfw, it's supposed to provide traffic shaping mechanisms, in 10.1, all those commands were listed in the "usage" but they didn't work. In 10.2, instead of adding the functionality to the kernel (it's a kernel issue), they just removed the commands from the usage, but they are still there in the man page (I'm talking about pipe and queue). Or another example:
man mount
.
SEE ALSO
mount_devfs(8)
% man -S8 mount_devfs
No manual entry for mount_devfs
Really, the whole BSD stuff is just one big hack. Some things work, a lot don't.
And here's my favourite:
Code:
% cat /sbin/mount_ftp
#!/usr/bin/perl -w
#print "this is a test:" . @ARGV . "\
";
#for (my $i = 0; $i < @ARGV; $i++) {
# print $ARGV[$i] . "\
";
#}
So, the guy who has written this script didn't know argument lists and had to test them?
Then it goes on:
$items = 0;
for my $arg (@ARGV) {
$items ++;
}
WTF? He could have used a scalar assignment or scalar(@ARGV) or $#ARGV +1 or at least a simple for (@ARGV) or foreach loop.
I mean, if guys who write this also write the kernel, how can this be a good system?
Do this experiment:
mv /etc/hostconfig /etc/_hostconfig
reboot
Have a nice day, the removal of a simple!!! config file prevents the system from booting.
(when you see the bash prompt, remount the root file system as read/write:
mount -u -o rw /
and then put the file back to its old place.)
Why do I even know this? Because a few days ago, something (I still don't know what exactly, but I suppose it was the sharing control pannel) moved the file to hostconfig.old.
It's those little things that often make me wonder. I really believe that darwin is not mature enough to be seriously considered a server OS. And even without those issues, I can't see a single reason why I should choose it over linux.
Finally:
4) The user base is too small. If you have an issue, nobody can help youl.
Just because the man pages aren't up to date and the user base is too small, it's a consumer OS?
We have absolutely no stability problems (8 servers running Apache, managing users and their homedirs, file server, mail server, eRez in Tomcat, JBoss and more), and I think you'll like the many new things in 10.3.
No reasons? Perhaps you can't see a reason, but there are many, and if you have Mac OS X clients, you'll be stupid not to use Mac OS X Server.
Originally posted by JLL
Just because the man pages aren't up to date and the user base is too small, it's a consumer OS?
You mean, because it isn't written in the man pages that the kernel isn't up do date?
What exactly makes an OS a server OS in your opinion? Availablity of server software? Ouch, soundly beaten by linux. Speed? Ouch, beaten by FreeBSD and Linux, reliability? Come on, you really want to debate that? Please tell me where you think darwin is technologically more advanced than Linux.
I really don't know how anybody seriously can think a system as young as OS X is a good server OS at the moment. It's certainly ok to manage user accounts, but that's about it in my opinion. Maybe 10.3 will be better, maybe 10.4 or 10.5, but until then I don't consider this an option. On the other hand, I don't think that linux is a consumer OS, it's also just not ready. Anyway, good luck with your servers.
Originally posted by 123
You mean, because it isn't written in the man pages that the kernel isn't up do date?
What exactly makes an OS a server OS in your opinion? Availablity of server software? Ouch, soundly beaten by linux. Speed? Ouch, beaten by FreeBSD and Linux, reliability? Come on, you really want to debate that? Please tell me where you think darwin is technologically more advanced than Linux.
I really don't know how anybody seriously can think a system as young as OS X is a good server OS at the moment. It's certainly ok to manage user accounts, but that's about it in my opinion. Maybe 10.3 will be better, maybe 10.4 or 10.5, but until then I don't consider this an option. On the other hand, I don't think that linux is a consumer OS, it's also just not ready. Anyway, good luck with your servers.
If you ask me, at this point of maturation OS X server is far more stable and useful than the Windows NT.
You seem to have very good skills as a Linux/Unix admin. And you are probably well compensated for. But these days its about $$$$'s, if I can get a OS X server to run on a cheap server and use it as a workgroup server. I wouldn't need someone like you. And save myself $100K a year. If something goes wrong, I just re-build the server from the Disk Image and be back up again in no time. Plus I don't have to deal with Geek head like yourself. (also I have Appleshare IP server, that keeps ticking along, without the benefit of UNIX or LINUX)
So I think one of the reason you should list is that, you want your job security.
I'm not a LINUX expert so I don't really want to get involved in that stuff. Just like a car mechanic wouldn't know anything about Jet engine, I don't want to deal with it. But for my Workgroup, I have to deal with Windows 2K using DAVE client. I think I could have saved the company money by just getting XServe, but thats where Intel bigotry comes in.
So if I could use one of the PowerEdge, it would have been good for me and the group.
I'm not talking about running FTP, Web or E-Mail system here, its just a Fileserver that PC and Mac users can use. NO command line involved, just turn it on, set it up and let it go. Let see LINUX do that. (Well I shouldn't say that because one of you will probably say you could do it while blind folded)
And some of you raised a question about OS X and OS X server being so similar that people would start to use it on cheap Intel box. And suddenly this will spread like wildfire among PC users.
Well I have one answer to that, most PC users aren't the experimental type, I'm not talking about Techheads here, I am talking about "My computer doesn't work, Turn it on? I just thought it was on all the time" type of people. And loss of revenue from these Techhead will be offset by the licensing fee that Apple will collect. 90% of PC users are people who can't use the computer in the first place.
Think about it for a second, these people who would use OS X on cheap intel box are not someone who buys Apple hardware in the first place, so they are not lost revenue to begin with. And Apple loyalist would never use Intel anyway. And I doubt any corporate IT manager will buy a Dell boxes and use a Hacked Server Turned Client OS.
And if Apple licensed the OS X Server to Dell, would Dell just do half-ass job and put it on the Market? I don't think so. If Dell get the license, you know pretty damn well that they will make that thing run smoothly on THEIR servers, so if they have any issues they will bring that up with Apple. So the Slickness of OS X Server will not be lost just because its been ported to Intel chips. Also I would think Apple would work with their Licensee to make sure it worked.
And who wants to type in that long ass serial number for OS X Server on each install.
Originally posted by 123
Er, the first one?
Seriously, OS X is a consumer OS.
Yeah you are right, OS X is a consumer OS, but OS X Server isn't.
Thats for people who want a server that works right out of the box with minimum effort. You don't have to fiddle with it.
Originally posted by 123
4) The user base is too small. If you have an issue, nobody can help youl.
This remark makes no sense what so ever. If its Software issue, Apple will help me with it, If its hardware, the manufacturer of the Server will help me. What are you talking about.
Does Dell not support their hardware if a drive fails or RAID controller goes up in smoke? and if a service on Windows 2K stops working, don't you contact Microsoft? I don't know what you are talking about, seriously what do you mean by this?
Originally posted by 123
What exactly makes an OS a server OS in your opinion? Availablity of server software? Ouch, soundly beaten by linux. Speed? Ouch, beaten by FreeBSD and Linux, reliability? Come on, you really want to debate that? Please tell me where you think darwin is technologically more advanced than Linux.
Jesus man, only the fastest OS and the OS with most server apps can be a server OS? There can be more than one server OS in the world you know
And if you want to know it, JBoss on our Xserves is much faster than a similar priced Dell running Red Hat (our developers test on that).
I haven't said a word about Darwin being more advanced than Linux, but if I wanted maturity and stability I wouldn't touch Linux - it would be BSD.
Again:
Originally posted by 123
I really don't know how anybody seriously can think a system as young as OS X is a good server OS at the moment.
As I said we run many things without any problems whatsoever and I don't want to pay twice as much for a Dell server to get at the same speed level in JBoss as we have now.
Originally posted by JLL
Jesus man, only the fastest OS and the OS with most server apps can be a server OS?
No, OS 9 can be a server OS, too. Enjoy.
Originally posted by 123
No, OS 9 can be a server OS, too. Enjoy.
Jerk!
You can't even give any reasons as to why Mac OS X Server isn't a server OS.
Man pages that aren't up to date and less server apps than Linux are laughable arguments (I don't even think that the number of socalled server apps are that different anyway).
Originally posted by JLL
Jerk!
Man pages that aren't up to date and less server apps than Linux are laughable arguments (I don't even think that the number of socalled server apps are that different anyway).
You know what is laughable? Running Java on OS X and think it is the fastest.
By the way, when was 1.4.2 released for OS X...? And how long did you have to wait for 1.4.1? (not even talking about the lacking official support)
If you don't want to realize that OS X was developed to replace OS 9 and was not written to be a top notch server OS, nobody can help you.
For the last time, my arguments:
- The BSD layer is a hack and you can see it everywhere (instability, commands that aren't supported by the kernel).
- OS X wasn't designed with speed in mind, nor with stability. It was the quickest solution when they needed a more modern OS for their OS 9 users.
- OS X was not designed to be a server OS, it wasn't even tested when they released it (you couldn't open more than 32 CGI processes on the first release of OS X Server without CRASHING the system : kernel panic, only you know how this is supposed to be a server OS)
- Apple isn't serious about the UNIX underpinnings. You laugh at man pages, of course, you never had to use them. But they are in fact important for developers. As is other developer documentation which is often non-existent when it comes closer to the kernel. It makes writing software much more difficult than it should be.
- Most software has been written for other platforms first, the OS X port is either not feature complete, is not as stable, is beta forever, is a version number behind the original release or whatever. Security patches are delayed etc.
There is no such thing as a Server OS, or a Consumer OS, there are no requirements that have to be met. BUT: based on other systems that are available, based on the features, the speed, the stability, the software situation, its (OS X's) original purpose, my own experience, Apple source code I've seen, conversations I've had with employees, Apple's marketing, Apple's history, bugs I've seen or heared of, the mailing lists etc., I think I can say that there are clearly better options to do most (server) jobs (managing OS X clients or other OS X interaction being an exception). Something I don't say about the client. That's why I say, OS X is not a server OS (you don't want to run it on your Dell server (what was it?), there are much better options), but a client OS (you want to use the Finder etc.). That's how I see OS X should be used.
Originally posted by 123
You know what is laughable? Running Java on OS X and think it is the fastest.
Well JBoss is the fastest when compared with similar priced x86 hardware. Get over it.
The more you write the more I think that you haven't tried Mac OS X Server. I think you're thinking about running swing based Java apps - I'm talking about J2EE.
A little thing as starting the app takes 2+ minutes on the developers' PowerEdge server running Red Hat.
On our Xserve it takes 30 seconds. Mac OS X's SMP capabilities are king of the hill here.
Originally posted by 123
For the last time, my arguments:
- The BSD layer is a hack and you can see it everywhere (instability, commands that aren't supported by the kernel).
- OS X wasn't designed with speed in mind, nor with stability. It was the quickest solution when they needed a more modern OS for their OS 9 users.
- OS X was not designed to be a server OS, it wasn't even tested when they released it (you couldn't open more than 32 CGI processes on the first release of OS X Server without CRASHING the system : kernel panic, only you know how this is supposed to be a server OS)
- Apple isn't serious about the UNIX underpinnings. You laugh at man pages, of course, you never had to use them. But they are in fact important for developers. As is other developer documentation which is often non-existent when it comes closer to the kernel. It makes writing software much more difficult than it should be.
- Most software has been written for other platforms first, the OS X port is either not feature complete, is not as stable, is beta forever, is a version number behind the original release or whatever. Security patches are delayed etc.
Blah blah blah blah. It works and we have no crashes. And the Mac OS X ports not being complete or a version behind? Not the server apps we use (JBoss, Apache, Tomcat, eRez and perhaps soon Oracle 10g on G5 Xserves).
The developer of the eRez Java app we use is very surprised about how well his app runs on Mac OS X Server, and since August 8 one of the two eRez servers has processed 3.7 million images.