Mac OSX Lion server?

[mrsir2009]

Will there be a Mac OSX Lion server? Because the changes (such as the launch screen) seem to be only handy with consumors, not professinals running servers. If a Lion server version doesn't get released, it will be another stride away from the professinal market for Apple

 

[roadbloc]

I seriously doubt there will be a server edition of Lion.

 

[0007776]

Yes there will be, They just like to show off the consumer features. But I think Apple sells too many Mac Mini Servers to discontinue that. The only reason they discontinued the xserve was because there were very few people buying them, and they probably did their research to find out that those people mostly could get by with mac pro or mini servers which still need OS X server.

 

[mrsir2009]

They are still selling it, but its going on January. Seems like Apple isn't making as much money off professinals as they used to.

If they do release a Lion Server, I'm wondering what is going to be different that will benifit servers.

 

[jaw04005]

Maybe they'll just lift the 10 connection limit and include the OS X Server tools on the retail install DVD.

There's not much difference between the two to warrant a separate SKU.

 

[MattInOz]

Didn't apple drop the 10 appleshare client edition for Snow Leopard.
And drop the price unlimited version.

I think there will product called Lion Server what form it takes is the interesting thing. There are places ServerOS works well like small design shops. I hear it's ok for testing services as well. Where it seems to fall down is sorts of big connected deployments that need hardware like the xServe.

I don't think Apple will kill Server Tools because they killed xServe hardware.

There are interesting ideas on how they could address the products future either a straight update, Appleinsider are suggesting a ServerAppStore. They need to address how to get it work better as a cluster, then again that is important for iOS and OS X.

Apple might be happy to give up the rack space but they want to be a device of choice for the admins and developers.

 

[BornAgainMac]

The Lion version of server could bring better integration with iOS devices. Right now even Microsoft Exchange works better than push email from Snow Leopard Server. A lot of improvement could be done with Mac Server with iOS integration. Plus some of the enterprise features that Microsoft server professionals have been asking about could be included.

 

[Winni]

I also doubt that there will be a server edition of OS X Lion. After dropping the development of their server hardware, the further development of an own server operating system just doesn't make sense anymore. Or will they allow the installation of OS X Server on third party hardware? I doubt it.

It's also common knowledge that Apple doesn't even use their own server OS and server hardware in their own data centers (they use Sun/Oracle platforms instead), so it's a safe bet to assume that almost nobody uses Apple's server products. So why keep on developing and supporting it?

I think that at this point it would make more sense for Apple to just drop OS X server and instead provide better integration and management tools for OS X clients for the Microsoft Windows Server family. That would actually make sense.

But then again, Apple nowadays is all about iOS and mobile Internet, and Apple's Mac unit (or whatever they call it) makes its money with home users, not business customers. Probably nobody would ever miss Apple's server tools and software if they just discontinued it.

 

[logandzwon]

I also doubt that there will be a server edition of OS X Lion. After dropping the development of their server hardware, the further development of an own server operating system just doesn't make sense anymore. Or will they allow the installation of OS X Server on third party hardware? I doubt it.

It's also common knowledge that Apple doesn't even use their own server OS and server hardware in their own data centers (they use Sun/Oracle platforms instead), so it's a safe bet to assume that almost nobody uses Apple's server products. So why keep on developing and supporting it?

I think that at this point it would make more sense for Apple to just drop OS X server and instead provide better integration and management tools for OS X clients for the Microsoft Windows Server family. That would actually make sense.

But then again, Apple nowadays is all about iOS and mobile Internet, and Apple's Mac unit (or whatever they call it) makes its money with home users, not business customers. Probably nobody would ever miss Apple's server tools and software if they just discontinued it.

I hate to be so directly confrontational, but you are way off-base here.
Apple uses their own servers. Judging by the job postings on their site they use OS X server almost exclusively in Cupertino. Going by the same source the N.Carolina postings speak heavily of Solaris. Now, this might mean the N. Carolina data center is a Solaris data center, however, they might simply be trying to higher people with Solaris experience because they are plentiful in that area. Solaris is not very different from OS X as far a SysAdmin goes. Either way, that project, while huge on it's own right, hardly qualifies as the majority of apple.

OS X server is very, very effective at the specific tasks it's designed to perform, (mostly groupware stuff.) It is extremely cheap. Really, it is just starting to become mature and small buisness are taking notice. The xserve was killed because it was unuseful in markets where OS X Server shines. I smaller office would prefer a tower form-factor, a tiny office would prefer a mac mini. Granted, they do need to update the ancient webmail client.

 

[Truffy]

Didn't apple drop the 10 appleshare client edition for Snow Leopard.
And drop the price unlimited version.

Yes, they did. The 10-client licence version was dropped, and the price of the unlimited version became that of the 10-client version.

 

[CrazyCleave]

Didn't Apple release a server version of every OS X release?, even Leopard to Snow Leopard? If so, it would be strange to not have a Lion Server. I'm an I.T. consultant in the SMB and Corporate space, and the day-to-day conversation I have with clients and colleagues is very contrary to what I read about Apple pulling back from the corporate world. I've supported Windows, and Mac for 15 years and the last year has been the most active for CIOs asking for Mac designs. The general sentiment is "Our CEO got a Mac at home and loves it. Were getting a lot of pressure to bring them into our network."

Obviously, this is an extention of the iPhone halo effect, and the challenges of getting Macs to meld with a Windows network still remain; but I've deployed more Macs in the last year than I have in the last decade. As far as Mac servers, including XServe, I get asked about it weekly.

From a practical standpoint, we deploy the Mini more often than the Pro or Xserve, even in rack environments. This is due to admins wanting to "dip their toes" into SL Server. They will set it up as a test machine, etc. I have a large, five state financial client that installed a Mini last month to serve streaming MP3s around the network.

I expect to see more and more servers deployed over the coming years. I don't know the specifics of why they dropped X Serve, but I have to think it has more to do with the form factor than the OS.

The biggest challenge my organization has with SL Server has more to do with virutalization and vertical database applications than anything else. But I do see more and more developers jumping into the Mac world. I never thought Quickbooks would develop a network/multi user version of QB Mac, and now we have it. I spent an afternoon calling old leads in the small business world that wanted to switch from Windows but couldn't let go of Quickbooks. That was a huge step forward.

The Exchange integration gets better and better, but imo it's still not ready to replace an Exchange server, just co-exist with one.

BTW, take a look at the Virginia Tech supercomputer (and others) to see how nicely the Mac Pro can fit in a rackspace

Cleave

 

[Anonymous Freak]

Didn't Apple release a server version of every OS X release?, even Leopard to Snow Leopard?

Yup, there were even "Mac OS X Server" releases years before the desktop releases. (They were versioned as "Mac OS X Server 1.0" through "Mac OS X Server 1.2", BEFORE "Mac OS X 10.0", and while the underlying "guts" were basically the same as now, they used a GUI that looked nearly identical to the older "Classic" Mac OS.)

 

[steviem]

They need to make it so businesses can host their own MobileMe infrastructure on OSX server.

If that is integrated into OSX server, some nice way to manage iOS devices and publish apps to users, then apple would probably have to come out with a new rack mount server (or allow osx server on vmware on rack mountable servers)

 

[SpaceMagic]

I really don't think there will be OS X Server. They will give GUI tools for server functions on the Retail DVD.

 

[Silencio]

No way will Apple discontinue Mac OS X Server. Not a chance.

As to what hardware it will run on, that's a good question. I think Apple sees the 1U form factor as being obsolete, and the Xserve was never a big seller anyway (not that Apple exactly went to the wall to give enterprise admins everything they wanted).

I'm an IT consultant myself, and the discontinuation of the Xserve has little to no impact on my near to mid-term planning for my clients. A couple of my clients have Xserves, but they have enough rack space to accommodate a Mac Pro in the future and don't really need dual power supplies, LOM, or anything like that. Most of my server installations are already tower systems because our racks are set up in narrow spaces that an Xserve won't fit into. Some of my clients with older Xserves will easily be handled by the new Mini.

I am curious to see if Apple either a) licenses Mac OS X Server to run directly on third party hardware, or b) licenses it to run in VMware. There remains the possibility of c) a revised Mac Pro that is more suitable for racking horizontally, I guess. And who knows what they have planned for that massive data center in North Carolina?

 

[CrazyCleave]

I am curious to see if Apple either a) licenses Mac OS X Server to run directly on third party hardware, or b) licenses it to run in VMware.

Both Leopard Server and Snow Leopard Server run as guest OS on VMware Fusion 3. Fusion 3 is also optimized to run with SL Server as a host. Perhaps you mean ESX?

Anyway, if you are VMware Certified, you can login to vmware.com and read a lot more about it, including the licensing requirement.

We run VMware all of our in-house 10.6 servers without issue (or license violation.)

Cleave

 

[deconstruct60]

I am curious to see if Apple either a) licenses Mac OS X Server to run directly on third party hardware,

Unlikely because this flys against there stated goals ( to deliver systems . not just software or not just hardware ).

or b) licenses it to run in VMware.

Do not need one. You can already run Mac OS X Server in Fusion ( so runs in one VMWare product).

You can run Mac OS X Server on raw hardware :

http://www.parallels.com/products/server/mac/baremetal/

just not from VMWare. It is purely a matter of VMWare ( and Citrix/Xensource ) deploying a hypervisor that doesn't allow Mac OS X Server guests to do live migration to other machines and validating the guest is about to start on a "Apple labled" box. for example if VMWare just got ESXi for Mac Pro and just didn't allow it to do the "advanced management" features from a Mac Pro.

Even without motion/live-migration a hardware hypervisor is useful. People often create cluster pods of similar virtualization boxes anyway. Apple would buy into a strict system that only allowed Apple-to-Apple migration would work for them also. There are limits to live migration anyway so this isn't a show stopper issue.


Part of the problem is a catch-22. VMWare isn't going to do a custom ESXi if there are no buyers. If not enough folks buy Mac Pro Servers then they may not ( unles perhaps Apple pays them for a custom ESXi embedded. ). The other half of the catch-22 is that folks aren't going to think to ask for a ESXi unless on Mac Pros if are fanatically fixated on the XServe.

the other problem is that most of the virtualization vendors give away just single box virtualization for free. ESXi and the entry level Xenserver are free. They change money only for the fancy stuff like live migration etc. Some of that is what Apple disallows. Perhaps a model where the virtualization vendor gets paid for every Mac Pro Server would solve that, but the fact that all of the commerical Linux distributions come with free (gratis ) KVM virtualization it is hard to charge money just for the basics. That is the catch 22 for Apple .... they can't leave out virtualization either. All the serious server OS have it too, for free (either in bundled hardware for AIX / PowerVM boxes or in the OS like Solaris Zones and Linux KVM ).

There remains the possibility of c) a revised Mac Pro that is more suitable for racking horizontally, I guess.

The handles on the Mac Pro make it about 2 inches taller. Minus 2 inches it would be 19". It wouldn't be hard to put the Mac Pro on a case diet and trade off perhaps a 2 inches from that 19" in height for 2 inches in depth. They could use exactly the same motherboard and most of the internals to do that (maybe sub in two smaller 700W supplies.) . Just like with the mini; small case mod ... ta da sever veresion.

And who knows what they have planned for that massive data center in North Carolina?

regular back end servers that other top 10 web data centers run.

Apple can run parts of their internal operations on Mac OS X Server. (servers for desktops , local DNS, Open Direcdtory ) But that doesn't mean you would want to run a web farm with them. There isn't really any real cost advantage in that since many of the other candidates (Linux , Solaris , etc) have been fined tuned (and apps tuned: java app servers ) to run that kind of load. Mac OS X Server isn't a big cluster ( computational , web ) OS.

Guess what HP, IBM, Oracle ... they've got Mac sprinked around several places inside the company. Nobody big runs 100% their own "dog food" . Microsoft maybe but that only because they only make half the equation (just the OS).

 

[deconstruct60]

Both Leopard Server and Snow Leopard Server run as guest OS on VMware Fusion 3. Fusion 3 is also optimized to run with SL Server as a host. Perhaps you mean ESX?

You have to be careful not to run anything on the host OS that would invoke overhead. It is easier to squeeze out more performance ( host more VMs ) on a machine with just a hypervisor. However, if the servers are very low I/O with low CPU and Memory overhead can get away with Fusion.

You also get many of the benefits of a LOM if the underlying host OS doesn't every go down. You can login into the host to inspect/restart/turn-off/etc. the guest OSes from over Ethernet. Again much easier to do if lock down the host OS so that it isn't doing anything other than running the guests and running the occasional VNC or sshd session. A Server OS worth its salt shouldn't spontaneously combust if just running two relatively well behaved apps.


It is getting to the point though where some folks say VMWare and mean running the entire kitchen sink of products virtualization infrastructure products.

 

[deconstruct60]

Will there be a Mac OSX Lion server? Because the changes (such as the launch screen) seem to be only handy with consumors, not professinals running servers.

There was almost nothing of Lion in that presentation that Apple did. You'd almost want to question whether there would be anything substantive in Lion.
The spent an chunck of time on "full screen apps". Geeee those already exist. ( lightroom, iLife , etc......) . the mac App store.... shipping at least 9 months before Lion does. New finder expose .... seriously?
It is another app update. There are other folks doing finder replacements

http://raskinformac.com/

where are the new hardcore foundation of the Lion in anything they present even for the client version. It was all shiny babbles. I understand they were presenting mainly to potential stockhodlers and not developers.... but like the old commerical goes.. "Where's the beef? " (http://www.youtube.com/watch?v=aISkVvi5iI8) Even for the client version. That presentation was mostly S.S. Big Bun.



The other wierd thing is that Lion isn't being shipped to developers yet. How long are folks going to get to test the new OS. A super short somewhat open beta program most likely means apple isn't doing much random testing for anything substantively new .... or is going to launch with major bugs.

 

[Cougarcat]

The other wierd thing is that Lion isn't being shipped to developers yet. How long are folks going to get to test the new OS. A super short somewhat open beta program most likely means apple isn't doing much random testing for anything substantively new .... or is going to launch with major bugs.

Relax. There's still plenty of time. The event wasn't developer oriented, it wouldn't have made sense to delve into the under-the-hood stuff like resolution independence or a new file system.

If Apple keeps on track for the summer, I'm sure they will have some sort of event for developers in March/April, with a beta released then.

 

[deconstruct60]

Relax. There's still plenty of time. The event wasn't developer oriented,

Then don't need an "event" to release a beta copy. The only thing an event does for developers is allow them to say they are working on updates for 10.7 in public. The snazzle and pop they would present is again primarily to sell stock and geared toward users. Again it isn't information being communicated to developers, it is primarily to allow them to talk about snazzle and pop to set some guidance for their users.

... it wouldn't have made sense to delve into the under-the-hood stuff like resolution independence or a new file system.
If Apple keeps on track for the summer, I'm sure they will have some sort of event for developers in March/April, with a beta released then.

That is an insanely short time for developers who have substantial apps to test on top of substantive changes while also addressing the issue of getting new features out for their users.

For example Apple ended up pulling the 10.6.5 release for server because they jacked up a couple of bundled server apps with that small point update. What is going to happen when they do a major point update from x.6 to x.7 ? They are going to break lots of stuff. At that point the app developers need to figure out if it was them or Apple's test are not sufficient ( or deviant from the specs ) . Extremely few programs as complicated as an operating system survive first contact with users without new bugs/defects surfacing.

The shorter the cycle the more likely it won't be complete or that app developers will have to advise their users not to upgrade. The shorter cycle makes alot more sense if there is very little to no new under the hood stuff. If it is yet another "refinement" release where they mainly just fix and minor clean up for the current stuff under the hood and/or attach purely optional "side cars" to the OS. The latter are typically new bling-bling APIs or just stuff not really part of the OS at all; just new bundled applications.

If it is an incremental refinement upgrade then keeping Mac OS X Server stack makes even more sense because it is mostly done anyway.

I fully understand there is a tension of handing out a beta too early and getting 100 different bug reports about a bug you already knew and documented but was missed by the developers because the Release Note's bug list has 100's of entries that they didn't wade though. However, the shorter the window the more the process turns into "If you don't find a show stopper, thoroughly embarrassing bug we're going to release this no matter what."

If Apple had "insanely great" testing then pulling minor point updates after distribution wouldn't happen. My suspicion is that Apple too overconfident in their test methodology. They'll hit their date because they'll just ship anyway.

things like "new file system" are a big deal. Resolution independence is probably going to jack up more than few apps because folks either had to kludge around it (and now need to remove the hacks) or push/pull screen elements into locations that app didn't expect them too.

Jan-Feb would be a short window. March is late. April is waaaaay too late [ given have to trim off at least a month for manufacture and probably another 1-2 months for a "bug freeze" state where anything new goes into the 10.7.1-2 update )

For example Red Hat Enterprise Linux went beta in April
http://press.redhat.com/2010/04/21/...x-6-beta-available-today-for-public-download/
Beta 2 in July
http://itmanagement.earthweb.com/os...-6-Hits-Beta-2-Targeting-Year-End-Release.htm
Probably targeting end of this year.



P.S. doing developer releases a couple months ahead time for iOS works because the apps are relatively small and immature. Lots of features and resources are on lockdown so can't get to them. The hardware interfacing with is highly homogeneous. Up until recently the screens are all the same size ( even with expansion not dealing with much variation. 3 is still quite small. )