srochford

The other thing to do is to force the date you write to the file to be in a "universal" format - something like: ofile.writeline year(now) & "-" & right("0" & month(now),2) & "-" & right("0" & day(now),2) & " " & right("0" & hour(now),2) & ":" & right("0" & minute(now),2) - gives things like "2008-07-11 12:10" which is easy to work with (you can insert it into spreadsheets and databases in any regional format and it will be "right")

The kids that are likely to cause trouble will not particularly be put off by the fact that the bar is not there. For example, they will probably have already spotted that when they go to open a file it shows something like "username on server\share" and will have tried exploring "server\share" by various different means. As kmount has said, provided your permissions are correct, they can't get anywhere they shouldn't get to.

We have a folder called "drivers" on the c: drive of our base image and in there go all the drivers we have. Sysprep.inf then contains: [unattended] OEMPnpDriversPath=drivers (note you don't include the c:\ bit) and as the machine restarts it picks up the drivers it needs. I wrote about WIM files earlier but this is one of the areas where they are fabulous - the laptops we're now buying need a raft of different drivers so all you do is mount the WIM file, copy in the drivers, unmount and save and you're good to go. With Ghost you'd have had to restore the image to a machine, boot it with Windows PE (or similar), copy the files, load the image back to the server and all this can take an hour compared to perhaps 10 minutes for editing the WIM

Key thing about sysprep is that it makes sure that you don't get two (or more!) machines with the same SID (security identifier - the thing which "really" identifies the machine to Windows). Most of the time having multiple machines with the same SID doesn't matter (and there are other ways of changing the SID) but if you don't change SID then it will come back and bite you at some point. Sysprep can handle the rename for you (at minimum it can give the machine a random name so you don't get errors at start up; may not matter but it means you can't have a "run-once" script which (say) connects to a network server and does some finishing off.

If you've got Server 2003 or Server 2008 and you've got no other imaging set up in place yet then WDS is excellent. As others have said, the WIM file format is brilliant (think of it as a big ZIP archive only much easier to work with!) and because there's no cost to using WDS (compared to things like Ghost or Acronis) it helps when you're strapped for cash. If you have Server 2008 then the multicasting is absolutely brilliant - unlike Ghost and others I've seen, there's no need to have every machine ready to start at the same time - you just set up your server, set up a multicast session and that's it. You PXE boot your PCs, choose to install the image and off it goes. We're currently pushing out an image which is about 10Gb uncompressed (4.2Gb WIM file) and it takes about 10 minutes to actually get the image on the machine (100Mbit switch; using 90+% of available bandwidth!) and this time stays constant as you add machines. The only reason I can see not to use it is if you already have something else in place or you're using Linux machines (or other non-Windows machines)

Agreed. I think google crawls the site more than once a day - it certainly picks up new entries very quickly. I've no idea how easy or hard it would be to change the way the search works on the forums but I suspect it's not easy and it will anyway be different from Google. I think there's more value in learning how to use Google well (because you can then search many sites using it) than getting a better search for just this site. Don't know if it would be worth adding a Google search box to the site?

When you turn on auditing on the file system you will get a *lot* of info; I think you're looking for event 562 for delete (see sample below from my event log) and you can filter on this. Event Type: Success Audit Event Source: Security Event Category: Object Access Event ID: 560 Date: 07/07/2008 Time: 17:15:17 User: CISNT\1973 Computer: PC051208 Description: Object Open: Object Server: Security Object Type: File Object Name: C:\temp\trans\Translator.sa.CAB Handle ID: 3576 Operation ID: {0,12812947} Process ID: 2788 Image File Name: C:\WINDOWS\explorer.exe Primary User Name: 1973 Primary Domain: CISNT Primary Logon ID: (0x0,0x6C227) Client User Name: - Client Domain: - Client Logon ID: - Accesses: DELETE ReadAttributes Privileges: - Restricted Sid Count: 0 For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.

sounds like you're winning :-) The 1.1Gb of files from Friday mean that you haven't had a successful backup since Friday - if that's normal (eg you only do a full backup on Monday night) then there's nothing to worry about. If you think you do a full backup every day then check - it hasn't worked since Friday :-( You can delete the web logs - they're just useful so you know who uses OWA but if you're not bothered then just wipe them; again, you need to leave "today's"

OK; plan B: bang heads until you get a properly working DHCP in place :-)

Pretty sure that WMI won't but VNC does. What are you actually wanting to do?

Check the make of motherboard (quick Google suggests it's an ASUS) and then check the motherboard web site for bios recovery - many modern motherboards seem to come with ways of recovering from a failed BIOS update.

Make the profile the way you want it to be as a test user - make sure you remove anything "personal" - eg drive mappings, printer mappings, outlook settings. Log off, log on as an admin user and then copy that profile to \\\netlogon\Default User - when a user without a profile logs on, they will pick up that profile. For your users with existing profiles, just setting a roaming profile in AD will cause their local profile to get copied back to the server and then used from then on. Getting Outlook right can be harder - ideally, what you do when you install Office is to create a transform file that specifies the name of the mail server etc. If you did this then a new Windows profile will just create the outlook profile; if you didn't then you can still make it happen by using a PRF file - google for that and you should find what's needed.

Do you have working machine startup scripts? If so, add the appropriate "netsh" command to the script. ("appropriate" means I'm afraid I've not got time to google at the moment but google for netsh set dns and you'll find it ...) If you don't have startup scripts and you don't have too many machines then you can use sysinternals psexec to run the netsh command on each remote machine - something like: psexec \\computer01 netsh set "local area connection" dns 10.0.0.1 When you've got that done, bang heads until you get DHCP in place - life is just too short to run a network of more than a few machines without it :-)

Shouldn't need to run a defrag to get the space back but I have an idea it's one of those things which happen overnight. check in "Program Files\Exchsrvr\MDBDATA" for *.log files. If they're older than your last backup then the backups are not working properly - do a manual backup and the logs should be deleted (you mustn't delete them manually because you will be well and truly knackered if anything goes wrong!) Check in \windows\system32\logfiles\w3svc1 (or other w3svc folders) - these are the web logs; if you don't clear them out then they will grow. Look in c:\windows for any $ntuninstallxxxxx folders - these can be deleted if you don't think you're going to want to uninstall any hotfixes, patches etc (but don't touch $hf_mig$) Does the server have a second drive? if so, can you move the pagefile there? This will need a reboot but is often a quick way to grab back a few gigabytes of space! Check in "\Program Files\Exchsrvr\.log" - this folder has logs for all email transactions. I think the default is to keep 14 days but it can be adjusted. You can delete all but today's log safely if you don't use them.

Turn on auditing on the filesystem for the folder above the one which is being moved (eg if on the server you have d:\groups\maths, d:\groups\science etc then I think you want to monitor d:\groups). You also need to turn on auditing in the local security settings of the server. When you do this, an event will be written to the security event log of the server but if this is a busy server you could be swamped in data.

Exchange logs the basics about all emails as text files to a share called \\server\server.log You can then parse these using logparser (free MS Download) and you can also dump that into (say) SQL server for use later. You've found all the other stuff which has been written. As I understand it, you can do what you want in terms of basic stats monitoring (how many messages in/out per day; what size, how many Mb per hour etc) but if you start personalising it then you need to take more care. You also can't just say "we are going to read all your email" but you can say things like "we log all access to email and may access it if requested to do so by appropriate authority"

If you want a group to be local admins on a machine I would do it with the "restricted groups" bit of group policy - computer config | windows settings | restricted groups. Add the group "administrators" and change the "members of this group" bit to include (eg) staff to add the staff group to local admins.

They are worth money for the copper content - if you have a scrap metal dealer near you then they will probably take batches of 100 or so.

I didn't know it (but then I don't use HP printers :-)) In general, if you can use standard tcp/ip to connect to the printer then I would always do this - the more stuff you add to your server, the more chance you have of something going wrong!

Yes This is a real "piece of string" type question :-) It depends on the spec of the machine. On a core 2 Duo with 2Gb RAM it takes <20 minutes from starting to install to ready to reboot. On a 1.7GHz Celeron with 256Mb of RAM it takes more than an hour. It does run in the background (standard yellow shield with updates being installed at the specified time - will prompt when it's ready to reboot in the normal way for updates which need a reboot and it won't reboot automatically if you tell it not to. No real impact on network - BITS is designed to be nice to the network; it could take days for all the machines to download the SP but it doesn't matter - the update client won't try to install it until it's completely downloaded. We started pushing it out to just a few machines then adding blocks of 50 and so on and it's now going out everywhere. So far, there have been no problems (what's the smiley for "touch wood" :-))

One thing you could do is set the config.xml to do a cache of the local installation source but not install (). That will give you a whole set of files under c:\msocache\all users - you can then copy this whole folder tree to each machine. Once it's completely copied then you can run setup from there - this is basically what things like SMS do; I think they use BITS to "trickle" the files in so that you don't trash your network bandwidth but you do get it all cached eventually. Once you start running setup, it's running from the c: drive and should run reasonably quickly. I've done basically this and it does use the MSP files in the updates folder.

How is the data stored? (SQL or similar database; some other type of database?) What else changes regularly apart from the data? I ask because this might affect what you need to back up regularly. We use Opera for finance - it stores data in (hundreds!) of DBF files and also creates PDF files and all sorts of other stuff. We have virtualised this and it works fine; we take a backup of the VHD overnight and also copies of all the files but if we couldn't virtualise it then taking backups more than once per day could be hard work. Our personnel system is based on SQL server. There are lots of other files but they're essentially programs, forms etc and only change when an upgrade happens. As long as the SQL database is backed up then you can easily move to another system (you keep a "warm" system up to date with the programs etc and make sure it's updated whenever you have an upgrade). If the main system fails it's just a case of restoring the last backup of the database and off you go (fingers crossed!) Have they said why they won't support a virtualised system? I would have thought that in California you could almost tell them that you will shop them for being "un-green" unless they start supporting virtualisation :-) (I do know there are reasons for not going virtual - Exchange 2007, isn't supported, for example)

One of the things I've been trying to do is make Open Source software available on Windows - Softgrid makes that very easy. This means that users can try things like Open Office, Gimp, Inkscape etc on their Windows machines. We've no plans to move to Linux just yet (unless we moved completely we wouldn't save on licensing - the FE Campus agreement is based on the number of staff, not the number of machines) but I think we ought to be able to offer it to people to try!

What are the w32time errors? What time is on the client PC? Windows will fail to do all sorts of things if time isn't synchronised properly. Try: w32tm /config /syncfromflags:domhier /update and then w32tm /resync to force it to use the domain time sources and to resync (could be that someone has changed it to use an external time server???) Is the DNS correct? Can't remember if gpresult gives any info about why gp is not applying. Are there any 1202 errors in the application event log? If so, run these 3 commands: esentutl /g %windir%\security\database\secedit.sdb esentutl /p /o %windir%\security\database\secedit.sdb gpupdate /force /boot - first 2 check/fix the database; third just forces gps to apply.

Does the teacher also have a desktop computer? If so, could that be used elsewhere and a laptop bought as their single computer? If they don't have a desktop then how are they expected to do their work? As localzuk says, it's not expensive compared to salary costs to buy a PC/laptop for each member of staff and any school which isn't doing that is not a good school (in my not so humble opinion :-))