srochford

Or, better, do them as web pages - I'd guess as many machines can read HTML as can read PDF and the file format is generally smaller so it loads more quickly. I *hate* web sites which are just a collection of PDFs!!

It may not save time but it would certainly make the process more accurate -having to go through manual mini-setup on 48 machines would not be easy and is prone to mistakes. If he can set up a script from either a boot CD or PXE Boot to update sysprep.inf then it's probably just a case of switch on, press F12 and then start the next one. At some later point he goes back, reboots all of them and off they go.

I think I posted a link to my website a while ago - there are a couple of scripts here which will mark an account as expired and then delete the mailbox, home folder and user account for expired accounts (we use a 2 stage process to try and avoid accidents ...) the logic should work anywhere but you will need to come up with a way of deciding which users can be removed (eg we look at the student records system to see if a student is still current - for us, students can start and leave at any time)

do you have a PXE boot ability (eg WDS)? If so, then you could script starting the machine from the network and copying in a new sysprep.inf Alternatively, you could burn a BartPE/Ultimate Boot CD, boot from that and then copy in the sysprep.inf

Came from me :-) It's vbscript - it'll run on any machine as long as you've got internet explorer and a PDF driver installed (actually, it'll print to the default printer but the original request was for PDF) as you say, save it as .vbs Create a text file c:\temp\urls.txt containing addresses of pages you want printing then open a command prompt and type to get the web pages loaded in turn and then printed.

Take a look at the windowsupdate.log file in c:\windows - it's pretty verbose but it might have something which says "Error xxxxx" which could get you started. Also, try stopping the updates service (net stop wuauserv) and BITS (net stop bits - may not be started) then deleting the whole folder c:\windows\softwaredistribution and the windowsupdate.log file. Restart the updates service and run wuauclt /detectnow to force a new download cycle.

Codeplex is very much what MS does! If you take the view (and I couldn't possibly comment) that MS is just about taking control of the planet then having a site like this is perfect - it means that you get everyone else to do the work of showing why you should develop for Windows (rather than any of those other nasty OS) You could also argue that it saves MS the hassle of trying to put together working code examples for any of their products (the stuff that used to be on MSDN) and just let other people do the work for you. If you think of MS as being "embrace, extend, extinguish" then you might think that they're trying to do a sourceforge type thing but, basically, Windows only.

You have the fix for this now but there are often things like this which you want do do on lots of machines. A technique I often use is to use regshot (free download) to snapshot the registry, make the change and then re-snapshot. This shows you to see what registry changes need to be made to your systems.

Never had to do this on Netware so I'm interested to learn. I wouldn't have expected you to need to do anything so I'm glad that you've found that to be the case :-) In general, most file systems seem to use some kind of "hidden" identifier which you can't change - in NTFS volumes with Windows, for example, it's the SID which links the user to security and you don't change this just because the "text" name changes.

Could you put printers outside the classrooms? If so, then you could set up printer pools so that (say) 3 rooms share 2 printers - the server will print to the first available machine. downside is that unless you turn on notification you end up looking at both printers to see where your job is but it does give you some resiliency (1 printer out of paper or toner just means that the other is used) If that's not generally possible, could you have a corridor printer assigned as second printer - default is printer in room with second printer the one outside. Look at group policy preferences (needs Vista or 2008 machine to administer but works on a server 2003/XP network) to assign printers to an OU (assuming your put PCs in OUs to match rooms) If the printing takes place from the server then it will log each job printed; logparser (free MS tool) can be used to scrape the event logs and stick the data in SQL for easy querying or there are various printer charging solutions available (you charge zero for each job but it just collects data about who's printed what)

Also having problems here and it seems to trash the install as well - can't now start the sql server service! Fortunately, this SQL instance is just for "play" so it doesn't matter but it's not good :-(

Do you want semi-permanent work or short term contracts? The latter pays better (but doesn't have anything much in the way of job security) You could look at the various IT recruitment agencies (there are some on here) or you could go to something like IT jobs, Developer jobs & Programmer job search: IT job board search - Computing Careers and see what they have. I don't have a lot of experience with agencies (other than those who constantly want to place staff with me) but I know some people find them good (but then some people like Marmite :-)) Good luck!

That makes sense. I'm guessing you could use either a PC running Windows with Internet Connection Sharing enabled or just a Linux box configured as a router (although can you connect the N95 to Linux??)

I'd hoped that someone else who uses DansGuardian would have posted by now but they haven't so I've done a bit of reading ... I can't find anything which describes what you want to do - there's a lot of stuff which doesn't use any authentication - but I don't think you can do it. Your clients will connect to a process on port 8080 which doesn't make any kind of challenge. You request that process to fetch a web page. Assuming that your request is not a "bad" request then the DansGuardian process then tries to fetch the page by requesting it from Squid. Squid then issues an NTLM challenge to the DansGuardian process. It doesn't know how to respond to it and has no way of passing it on to the original client so it just sits there. Squid says "no valid response" and abandons the request. Your second option looks possible (but I'm afraid I don't understand IPTables either :-() - this time, the client will talk to Squid and respond to the challenge. If the client has a valid username/password then Squid will pass the request on to DansGuardian which then checks to see if the URL is OK and fetches it. It's certainly possible to get Squid to talk to another proxy - I've done it when the two proxies are on different PCs but I can't see why it wouldn't work on 1 PC - just need to get the ports right good luck!

Not sure what you're doing here, but if the data could already exist then this will either end up with duplicate records (bad :-() or just fail. What I always do is try to delete the record based on key field and then do an insert.

That's right - what I should have added was that possibly the mtedit program triggers this process so that clients redeploy and they get updated to point to the new location. Part of the reason for using MSIs is that they are "self healing" but this only works if they can find the original source. I suspect MS don't want you changing the source because then you break the healing bit. What I would have thought they could do, however, is add something which doesn't do a complete redeploy but just updates the data on the client machine to point to the new source.

I'm guessing the contacts were previously in "personal address book" rather than "outlook address book" (assuming you were using Exchange) if so, you need to log into Outlook and drag the contacts from the address book to the contacts folder. Alternatively, if you were using Outlook with PST files and are now using OWA to talk directly to exchange then the contacts are in the PST. Again, you can log into Outlook, open the PST and drag the contacts from the PST to the Outlook account. If they are in a PST then you may be able to automate the import using exmerge.

I think you also need to edit the .aas file to make the change - it's possible that mtedit is also doing this. Clients will then "see" that the package has changed and redeploy. I've never found a way to deal with this - it's why we now deploy everything from a DFS share. When you need a new server you add another replica, wait for it to fully replicate and then just remove the old server. If you don't want to use DFS you could also use a DNS alias on the server - then when your server changes you just update DNS.

Not sure that it proves anything except that some IT people are poor. I've spoken to people working in schools which are not BSF and they tell similar tales. What I think we should be doing here is publishing examples of how IT can/should/does work in a well run school and trying to make sure that those are used as the benchmark for any new provision. It's not enough for people to tell us "well, BSF is better than the worst schools now"; they ought to be saying "BSF will be as good as or better than the best schools now" Just as an example, a couple of the slides show "my profile is 5Gb" etc but exactly that point has been made on here recently as if it's the fault of the users - it's not. A well run network will not allow profiles to grow stupidly big and will be monitoring things like that to make sure that they're OK.

Videk is a good place to start for anything like this. Fans are here - the 1U general purpose one might be best (because it will fit any cab) Fabulous company, excellent service, not bad prices!

I don't use Dan's Guardian so I may be talking rubbish here :-) It looks as if you're doing authentication to Squid and then trying to pass that on to the Dan's Guardian process. If that is the case, then I'm pretty sure you can't - NTLM credentials won't do a 2 hop process.

Remember, the DC will actually do very little in the way of DNS lookups for itself. It will service lots of requests if it's hosting the DNS role but for itself it just needs to find info about where the other DCs etc are. Once that info has been found it will be cached. Even if it did need lots of lookups, DNS is a very lightweight protocol (watch it with Wireshark) so it's not going to mess up the network.

Agree with @Cookie_Monster - I've always known this to be the way to do it. If you have 1 DC then you point it at itself (don't leave it blank; set it to 127.0.0.1 - this is what will happen when you set up a new domain and let the wizard configure DNS for you) If you have 2 then point them at each other for primary and themselves for secondary. When you have lots, point them at another DC on the same site for primary and a DC on another site for secondary. I think DNS island problems were mostly cleared up with Server 2003 but we had that problem years ago on 2000 Server and it was hard work getting it fixed!

If you don't know the answer then the question isn't thick - and it's way better that you ask the question than just blunder into something :-) The time you do need more than 1 IP address is if you want more than 1 mail server. If you have multiple web servers then you can have 1 IP address and different ports and you just tell people to use the correct port (because you publish a link to it with the port). You can't do this with SMTP because incoming mail always connects to port 25. For example, we have cnwl.ac.uk and student.cnwl.ac.uk domains - these are handled by different boxes and they have different IPs (actually, the first one is now handled externally for spam filtering but it used not to be like that :-))

Why? DNS and WINS servers must have fixed IPs but while it might be "tidier" to use fixed IPs on other servers I can't think of any reason why it should be so.