powdarrmonkey

I have a power switch for that.

Forget the slow logon times, it was a red herring. The problem is that your DNS domain is school.com which, unless your name is "Office Depot, Inc." is not registered to you on the global Internet, nor do you control it. Within your network, that's only really a problem if you want to visit School Supplies: Teachers & Classroom Supplies at Office Depot (or any other related address), since it'll resolve (or not) to an internal address, which will probably tell you to go away. However, it's a significant vulnerability because according to the rest of the world, you don't control it. One example: imagine a laptop taken outside and connected to the internet at home: 1. Internet Exploder tries to do automatic proxy detection and requests wpad.school.com (WPAD) 2. a rogue administrator at Office Depot, Inc notices your DNS requests and decides to make it resolve to a server, so your browser requests /wpad.dat from it 3. rogue configures the browser with /wpad.dat to use his proxy server, and IE happily fetches it and acts on the contents 4. now all your traffic belongs to him, including passwords, credit card details, etc. User doesn't notice because, let's face it, they don't Seems innocuous enough until you look at it like that.

There's something else you've missed then: if you have two lines, you'll be liable for two duties. (note: these are only proposals.)

No, you're thinking the problem backwards. Any resolution of school.com inside will only ever return local records, not the ones published by the legitimate registrant of the domain. It reminds me of the problem faced by an organisation I shall not name who just picked a range of public IPs for their internal scope, instead of requesting a proper range from IANA or using a private range. Then they leased a handful of dedicated servers in a datacentre, and by complete chance they were allocated the same range as they had picked for internal machines. Result: had to totally redesign their internal range because they couldn't route to their shiny new (and very expensive) dedicated rack.

There are no 'backup' domain controllers any more, only domain controllers and domain controllers with extra perks. (all DCs are equal, but some DCs are more equal than others...) Anyway.. pick a server, any server. Most DCs are pretty lightly loaded, you don't need anything special.

do you realise how wrong it is to use a valid second-level domain that you don't own as an internal name?

/me used to, but hasn't been in years.

But if you've only got one UPS anyway, that hardly matters. If you've got one UPS and redundant PSUs, I say make the most of the PSUs. I suppose you could argue that with one UPS, the safest bet is to plug one PSU direct into the mains. Then if there's a power outage, you've got UPS, and if the UPS goes you've got raw mains. Bit screwed if the UPS blows and takes out your MCBs though.

False; any self-respecting Netgear switch these days has a 48Gb/s backplane, so fibre-fibre connections go at full pelt. However, I agree that if there's only one uplink to the switch, choose a 48 port unless you think you might want to separate them later.

It does.

And if you have a cable running to the house but don't have it connected, you'll still be liable for duty. For example, you move into a house that used to have a line, but you don't ever get it connected.

Not if a PSU blows up, a la Optiplex.

Er... well, kind of. Switchgear keeps a table of mappings from IP to mac for the purpose of delivering a packet, but there is no routing information. If you have the CCNA stuff handy, revise the OSI Model and remember that on an IP network packets travel from layer 7 downwards on the way out, and upwards on the way in. You can't miss a layer out.

No. You can't route packets on an IP network without an IP address to target.

If you asked for directions from San Francisco to London it used to tell you to swim across the Atlantic Ocean... disabled now

I thought all the GX series had internal speakers, in which case you just need the driver.

I still say try it and see - then we know, one way or the other instead of just conjecturing. Either way they certainly won't support it...

No, Mono is binary-compatible. However, the implementation is not complete, which will be the sticking point.

Not true. Mono is a free implementation of the .Net API; .Net itself is designed to be architecture-independent - like Java, but sensible. Ish. Anyway, JoshJohnson: your mileage will probably vary, but I'd have a go and see if it works. You might find you need WINE.

It's not unreasonable to ask them if this is suitable, be told that it is, and believe that in good faith. If he just assumed then I have no sympathy, but as he took the trouble to find out...

Use the PayPal transaction ID, which is guaranteed unique.

welcome to yesterday.

Hmm, dunno. It's worked for me everywhere I've done it.

There's nothing technically wrong with double NATting, it just takes some careful planning. Sounds like you need to talk money with them. It can work wonders if they think you might start being more independent.

Internally, it's to do with the new application separation model. Applications can be instance-aware in 7 as well, but it's done differently and they have to be updated accordingly.