I just shelled out on a mini server and a promise 8TB. My migration from 10.5 to 10.6 had gone extremely smoothly, so I was expecting nothing less. Boy, was I wrong.
Not only failed the old server to mount in target disk mode, the migration from the Time Machine backup failed the first time and kind of completed the second time. Nevertheless, all the promises Apple has made in their documentation are for someone else, but not for me. I couldn't get certificates to work because of some unspecified error (specifically said, that User intervention was not possible!). All other input was greyed out. I finally managed to get something done via the certificate assistant - not sure it works though as access to the mail server was still not working (STARTTLS was the smtp message incoming mail received) and users couldn't connect.
My server also acts as a router and does NAT b/w the outside world and the internal network. I work with static IPs and have about 60 or so addresses. Whereas in SLS you could define your internal network with whatever number combination you wanted (10.0.xxx or 192.168.xxx.xxx, etc.), I only found out after about three more hours, that LS only allows an internal network of the type 192.168.2.1 (see Gateway Assistant). Now, how stupid is this?
Another thing which I discovered: whereas in SLS, everything is neatly controlled in Server Admin, LS does away with this. NAT is set in Server Admin, but the detailed configuration is then to be found in System Preferences -> Sharing -> Internet Sharing. Somehow, Apple managed to tear the whole management process apart, with some things being specified in Server, some in System Preferences and some residual settings in Server Admin.
Server Admin gave you a lot of granularity about specifics of file sharing and the permission levels, introducing Access Control Lists, etc. Again, splattered about a number of different locations, one can only guess how to get this working again. I have a number of 'custom' settings for access. But there is no way to view or edit them sufficiently.
My verdict for the day is very bleak. If Steve had been more involved, the person heading this piece of software would not work at Apple anymore. It's a disgrace.
I am sitting now on a nice, but useless machine with a huge 8TB Thunderbolt drive.
My old SLS server has gone back online and is performing nicely.
I might go down the route of installing Parallels and running SLS in Virutalisation format as suggested in a previous post.
But for today, I am just shaking my head at what Apple have done. The signs are not good. First Final Cut Pro X being a complete disaster, now LS, what next????
One more thing: I think I will start another thread about alternative server environments for people frustrated with LS...
Sounds like my tale... I bought the Mac Mini server as well to replace an aging mail server and found Lion Server a massive mess (see my posting last page).
Just a comment on your posting. I don't think that Apple with a Steve Jobs would be any different right now. I am positive Steve Jobs was the one who set in motion the slow phase-out of Apple's psuedo-enterprise product so that they can focus 100% on being what they feel they do best: be the greatest consumer electronics/software company in the world.
I've finally faced the facts that the next version of OS X Server will basically never be released or will be much the same as LS, and that Lion Server represents the last step of the phase-out. I will probably go to some sort of Unix box solution down the road when my Xserves (3 of them) and mini mail servers (running SurgeMail now) die. That will be a sad time, but I am certainly not moving to Windows.
It does baffle the mind though how we got where we are. With Apple's financial position, they could easily have invested 100 or 200 million back into the server products line (mere pocket change) and just kept on improving OS X Leopard Server and slowly made deeper and deeper inroads into the Enterprise. With so many devices making it into corporations (iPhones, iPad, laptops, etc.) you would really think that Apple would have changed direction and spent "more" on enterprise rather than shutting it down. This decision could easily come back to bite them in the butt when IT admins start phasing out MacBooks, Airs, and iPads from the enterprise, because the Server software is too thin in features and capabilities, and too difficult to use to manage hundreds of users... Think about it...
I don't have any experience with those tools, but in the past I have used the Client verison of OSX and built it into a server. I have used Webmin for this with pretty good results.
I do currently have an xServe running snow leopard server, but the Lion server thing is a bust, as the migration tool doesn't work at all, and I have NO interest in spending days manually reconfiguring it.
So I guess the next time I need to "upgrade" I will go back to using the client OS and building the additional tools I need from the command line. NOT very Apple like to be sure.
See my one last thing thread: https://discussions.apple.com/thread/3399811
I think virtualisation is the way forward, i.e. run Parallels or VM Ware on a Lion machine (maybe even server, not 100% sure about the licensing issues) and then have a combination of Lion machines (see suggestion from Martin Joseph) or e.g. Ubuntu servers (DigiAngel) take care of the individual tasks.
Only problem with virtualisation is that it will only buy you 1-2 years before you have to abandon Snow Leopard. If Apple will follow the normal pattern, as soon as 10.8 is out, they will stop the security patches for 10.6 (only keeping the last one - 10.7 - patched. ) Since an unpatched system is clearly no way to go, that's the end of this road.
For myself, I'm happy with 10.6.8 for now. If my XServer dies, I will have it repaired or replaced with MacPros from the pre Lion batch. By the time this becomes troublesome, either Apple has released a real Lion Server, or we will have moved to a different OS. Evaluation starts now.
The only point which I still like to understand, is if Apple broke all the functionality simply by over-doing the effort to make it even simpler to use. Braking a working server on purpose with arbitrary limits like one sub-net 192.168.2.x only , one time machine volume only etc. does simply not make any sense at all. Unless it's the Home-edition which you want to price below the real server.
It is hopefully seen as a bug rather an improvement - and since I expect Apple to monitor these discussion boards, they should by now be aware of the issues we have. Since 198x I did never need to abandon the Apple OS so far. Even in quite some dark times. Let's stay optimistic for a little bit longer.
You can have more than one disk for Time Machine, simply use disk utility to make it a raid device. MySQL, Tomcat, etc. are easily installed via Macports, and, easily started up and stopped with one simple command, even at startup.
I am using LS just fine, but, I am a long time system admin on many platforms so command line doesn't scare me off. I do agree it's been WAY oversimplified though. Moves a lot of customers from being able to use it, to not. We use most of the functions of LS, but not NAT.
Ad RAID. Again, this works only if you have a rather simple set-up. We use currently 2 RAID crates, each with 16 disks inside. Connected via fibre channels to a xserver. Both crates have their own disk sizees, hot swapping spare disks etc. And we plan to expand.
I still do not see the point of crippling server software just to make it even simpler to use on the costs of functionality. Apple used to be able to combine simplicity and functions in a great way. So why stop now ?
If there is indeed a market for home user - fine - bring out a home server edition. Or an (non)expert mode in the server software. Keep charging more for the pro edition, if needed.
But in the current form of 10.7, we will have to start moving to new server solutions. After all these years I fought for the mac server (sometimes against the own IT department) I find this disappointing.