morganw

I've used a 20 unit iPod Touch dock that should sync and charge all of the units at once. In actual use it just demonstrated that iTunes and the Apple Mobile Device service are not designed to have more than a few devices connected at once. processor utilisation for the service would increase with every additional unit connected, in my case everything started to grind to a halt after about 15 iPods being connected, 18 iPods would lock up the entire system with the Apple Mobile Device service completely maxing out one processor core. I wouldn't recommend a dock for anything but charging. Hopefully iOS5 will allow people to self-manage over wireless to the degree they need.

I bought 5 upgrade licenses for the network activation version last year, got a price from them and placed an order. I then had someone else trying to convince me that the minimum purchase amount had to be 10 licenses and I couldn't buy 5. I didn't do anything as the purchase order had already been sent, the original order was processed and the license server works fine with a 5 licence key.

As drive sizes get bigger there is an ever increasing risk that array rebuilds will fail if a drive encounters an unrecoverable read error. To minimise the risk in RAID 5, use more drives that have a smaller capacity rather than a few big drives. If you want RAID 1 then the controller will be cheaper, and you could use multiple discs to keep multiple mirrors if it's a critical system (and contrary to what is written above it doesn't go offline when a drive fails) so I would say that it's more robust than RAID 5 but RAID 5 is more flexible as the controller should allow you to grow it with additional disks. If you just want reliability and aren't worried about speed or expandability then RAID 1 is safer. e.g. 4 x 1TB drives in RAID 1 gives you 1TB of storage that can take 3 drive failures.

I think the problem is that the way the program is used in schools requires a lot of file trading both internally and externally. Having it as a Windows only programs makes this process a lot harder for everyone involved.

I've been using Taskmagic version 1 and 3 and have issues running a silent unattended install through our usual deployment systems. By creating another process for the extractor you will mess up timings and dependencies, the deployment system thinks the program has finished installing when it hasn't because it isn't aware of the file extraction. In the end I've hard coded a timer to wait for 3 minutes to make sure that everything goes through. By using silent self extractors you also risk triggering anti-virus events if some malicious code decides to start using the same extraction techniques that you are using. Add to this that comdlg32.ocx is constantly unregistering itself when Windows Updates install so the program can't run correctly most of the time anyway. I'll just echo the above comments, I really think you should update your development environment. The program is worth the effort, you can tell that when you see staff and students using it, but there are technical factors dragging it down. If you had this working in a web browser and could demo it to some people on an iPad you would make a fortune.

If you look at the AP logs on the controller then you'll probably find that the channels are being changed or power levels are being adjusted too frequently. The system that I had installed a couple of years ago used to be fairly stable but recently degenerated into a mess after adding a couple more APs. I found that the automatic power and channel settings had everything going in circles and settings were being changed every 10 minutes. Try manually setting the channel and power on the APs that you are using and then give it another test.

Is the start menu redirected?

Old hardware should be okay, but the restriction is necessary for new hardware that will be marketed as Windows 8 Certified. I imagine if you are selling OEM Windows 8 licences you will only be able to do so on such hardware as part of the agreement.

Keys as in, code signing with a public and private key pair. The private key of what you are installing (i.e. the bootloader) will have to match with a public key that is on the motherboard.

There wouldn't be anything to stop a consumer key being supplied with hardware that would allow someone to do what they wanted, but I wouldn't be surprised if manufacturers refused to manage and distribute the keys (and I imagine Microsoft will create incentives for this to be the case). If you built your own PC then you will have to be given a key, the real problem will be for mobile devices as you can't really buy the parts and make your own.

They are my current first choice for everything based on customer support (although I've only ever dealt with one person there), they also did an EES and Adobe volume license deal for me and the prices were good.

For something easier than LFS you might want to look at SuseStudio.

You can do it with a local install of Thinstation that will connect to wireless.

I've had Debian running on a Microserver for a while now running Netatalk, mt-daapd, and Exim. Not had a problem so far, depends what you are looking to get out of it I guess. If you want to play around and learn things rather than just run stuff then it probably makes more sense to install a VM platform so you've got some rollback options.

I've got a very similar setup and exactly the same speed issues. I think I'll try deleting the stubpath key from the mandatory profile at the weekend and see what happens. The only thing I know for sure is that trying to map printers with group policy preferences makes it worse (adds about 30 seconds), so I've had to go back to vbscript for printer mapping.

.NET framework 4 broke my domain join but I think that was a timing thing, manually restarting the service would always make it work. I would remake your image in a VM so you can take disk snaphosts, then start with an empty image and see if that works. If it does then gradually add-in the rest of your software and see what breaks it.

My hostnames are specified in DeployStudio and get overwritten by the reverse DNS values.

Try temporarily putting the domain admin account credentials into the AD settings and give that a try.

If you restart the fog service after this error do you get the same error in the log file the next time around?

And you did put the password back through FOGCrypt using the default key now that you have gone back to the default hostnamechanger.dll?

And if you look at the host settings it defintely lists the username in the format "domain\user" for each host, rather than looking at the default values in the FOG AD settings?

At face value you would think that was due to incorrect ip or dns configuration, but you are obvisouly contacting the FOG server with no issues. I assume that manual domain join will work with no issues?

I've never seen anything like that in the log file before, I think you are right about the hostnamechange.dll. Just to for testing you could revert to the one that ships as standard and encrypt using the standard key. To be honest when I was running this I just created a new service account and delegated permissions to over computer objects in certain OUs, so if someone did decrypt the password they wouldn't have control of a domain admin account.

OS X will attempt to set an initial hostname by using a reverse DNS lookup on the current IP address. If you have short DHCP leases or the workstation name has changed then things can get a bit hit and miss. If you used DHCP reservations you could probably join a domain with right name every time, otherwise I would suggest always doing the domain join manually so the hostname can be checked.

FOG doesn't use netdom anymore, I don't think it has needed this for the last three or four releases. Also the newest version of FOG will attempt to change your sysprep.inf file to insert the hostname and domain joining options directly into the file, you can revert to the old behaviour by adjusting the settings in the UI. I've also had problems with .NET framework 4, in particular it would sometimes stop snap-ins from installing and hang before joining the domain. I've you got the time I would reinstall 0.32 from scratch in a test VM or similar just to see if you get the same errors, it sounds like something isn't right with the service. Are you sure you are using the latest version, you should download the latest through the web interface once you've run a server upgrade. Check what you are using against http://[serverip]/fog/client/