PiqueABoo

Well spotted - I've got a tick in that box too and it's certainly a credible potential cause.

::sigh:: Since we're now trading the obvious I don't mean when it was something that wasn't a credible MIS running on a Commodore Pet or whatever, at some point school MIS became something with a network architecture, multiple users and got one of those RDBs at the back end. No I don't know the nitty gritty history, but a quick Google says SIMs was turning over significantly into 7 figure numbers by 1995 which coincidently is when I reckon most of the IT world, especially the dev part was thinking very hard about how they would adapt their products to fit the Internet (YMMV but by the end of the decade I was seriously bored with those profligate web-interfaces that had managed to surface in every other application). In MIS world specifically, it's also the time when at least some of us were contemplating clunky Edifact files (or similar) and wondering about doing that with *more* standard (network) messaging APIs. Plus SIFA say they (or at least the educational sector wheeler/dealers) have been on the interoperability case it since 1997. It's 2010. Did you read my questions in the previous post, as in what gave you the impression I'm oblivious to that? SIF are apparently asserting that horizontal interoperability needs ZIS same as vertical operability. I'm asking, again, why that is the case? A lot of apps wanting to read a list a people is obvious and it might be nice to read some bits from places besides the MIS, but what is all this *bi-directional* data flying about? Why do we need to add more complexity and a point of failure i.e. a ZIS, when we seem to get by OK now by adding a single client account to any app that another wants to read from? And we're talking a handful of those other apps not 100s. A standard open protocol for talking to things in classic-client server mode is fine, but for horizontal I really don't get the need for the ZIS... or perhaps that should be ZISs plural, coz if you're building in a single-point of failure by design then I'd want to be replicating that ZIS function somewhere. I'm prepared to change my mind, but only if someone can give me some real examples of things like apps getting ASAP notifications of stuff happening in other apps being useful enough to justify the additional infrastructure.

Why it is perceived as a sterling solution for interoperability between internals app, as opposed to interoperability with external orgs (well assuming it is a good solution for that). IMNSHO had school MIS been done "right" in the first place they'd have looked around the wider industry a bit, thus factored in potential interoperability from the start and used that abstraction layer localzuk keeps on plugging... on one side the DB/whatever, in the middle access control + some of the biz-logic + whatever fits this variant of MIS, and on the other side a client messaging API... and with that adding another tarted up subset of that messaging API to the repertoire i.e. SIF or anything else, wouldn't be that big a deal. Me too, but can my evening scheduled AD user updater app do the obvious and go talk directly to the MIS via SIF to get a bunch of pupil/staff deltas since a given timestamp, or does it have to bounce that through a ZIS. Apps responding to a common messaging API is fair enough, but I'm really struggling to see why a) any app within an organisation would need or want another app to write to it, b) why an app would need to be quickly notified of changes made in another one, c) why you'd need a independent ZIS traffic-cop function to police relatively simple internal interoperability. That easy huh? Something has to run round the systems getting "all the information".. what is it?

My server and workstation McAfee "On-Access Scanner" properties has *no* tick in the box for "On network drives". And there's no sign of any related trouble on the workstations.

So to talk to a MIS somewhere else in the rack or even running on the same OS you need: a) Your app to talk SIF... fair enough, we have to talk something. b) A ZIS configured and running somewhere.. what resources and supporting apps does that need to run? c) Your MIS to talk SIF via some nearby "agent".. what resources and supporting apps does that need to run? Do you get those included for nowt or are they yet another cost option coz somehow all those millions just don't quite cover the dev costs. d) The MIS. Perhaps I'm way off but its got that multi-box (or at least VM) feel, and if it's something local wanting to track students e.g. library system/whatever, that just feels wrong. Does SIF mandate a ZIS for everything i.e. can you bypass that for local stuff?

An update re. my server: It's not conclusive, but there aren't any 2012s in the logs until after AV was installed on server and clients (there was a lot of data flying about both before and afterwards). I usually get a set of 6-8 at the same time (resolution = seconds), but there are a few cases where there are just two.

Thanks for your updates on progress with this. Courtesy of a few little glitches here and there, and some of that nasty old incompatible educational software, I wasn't entirely happy with Win7 and I've been cautiously waiting the SP1 before putting it into production. But I will be using all the ingredients mentioned on this thread.. folder redirection, folder quotas, offline files, IPv6 disabled same as I do server-side now etc., so I'm also interested in the outcome. Anyway, this update is interesting. I don't see why desktop.ini files would have anything to do with this, but offline files did seem like a vaguely credible contender for the deletion i.e. I could imagine that getting very confused and emptying the folder. -- Re. desktop.ini. I make a home drive mapped as H: e.g. ..\share\FredB But I redirect docs to a sub-folder ..\share\FredB\Documents Thus I don't care what desktop.ini tells Explorer to display. Note: I also hide that H: drive - two ways to get to roughly the same place is too challenging for some folk, so they just get My Docs plus those usual shared-with-other-folk drive mappings. Would have done without H: drive altogether, but some other things didn't work with a UNC.

It's an extremely serious one and I'd be interested to hear what you think is required now that you know we didn't have then (and based on what experience). Technical history did not start this century, IMO there hasn't been that much innovation for long while, there is however a lot of ignorance and wheel-reinvention, and 20 years ago does NOT mean fewer requirements than a modern school MIS. The only thing that's especially new is that it apparently takes the school MIS sector eons to sort out interoperability.. not that standard protocols were especially quick back then, but they were definitely a lot quicker.

We got by OK without relational before e.g. I first worked on a bigger, badder, distributed, interoperable MIS system 20 years ago that got by just fine without a proper RDB, and at a glance it fits onto the NoSQL/NoRel category, as apparently does one of IBM's from 1966. That system did eventually go RDB, but not for any pressing technical reasons i.e. the customer's consultants taught them to ask for it on tenders and you had to tick boxes. Naturally performance & resource usage went right through the floor so customers then had to pay for that tickbox by throwing much, much more significant hardware at it.

There are APIs and there are APIs. For instance, would it be implemented as a dotNet library or something more arcane/geeky? Mmm.. my off-the-cuff multi-tier architecture I didn't bother posting several eons ago in this thread wouldn't do that either. I guess you have to have been in s/w dev to appreciate this: Smart people who work in s/w can always point at truckloads of stuff that could be done better.. a lot of software has evolved organically, slowly over time which is punctuated by rushed deadlines... you end up with code containing naff compromises, legacy support, new bits patched on that don't really suit the underlying architecture.

Mmm.. saw this earlier.. new remotefx & hyper-v memory blah blah blah.. but what does it fix Microsoft?!?

Windows 32bit On Windows 64-bit = It's the (counter-intuitive) WOW64 that runs 32bit.

No hyper-v involved, but that large send offload coupled with another change I forget was someone's solution somewhere for this event on a hyper-v guest. + I've looked at a few and so far for me they're all the Invalid Device State error code 0xc0000184

Back to the drawing board (those offload changes didn't change anything for me). Can only think that something I'm doing must be causing these - I've got a new 2K8R2 box without AV on it yet, so if these 2012s haven't already started I'll watch out for them when AV goes on towards the end of next week.

But half the point of thinstation and the like is to just PXE boot the small image across the network i.e. you only need to go to one place when you want to update it.

Good thought, but which TCP/IP offloading? Looking at a system that produces most of these events (most workstations & the busiest) I've just gone to the Intel 82574L Gigabit NIC advanced config and turned off: - IPv4 Checksum Offload - Large Send Offload (IPv4) - TCP Checksum Offload (IPv4) IIRC the TCP checksum first has crossed my path in the past on some 2003 boxes, the large send was part of someone else's fix for a VM scenario but I'm not fussy, and I added the IPv4 Checksum because I suppose I may as well knock out IP checksums as well as the TCP ones. PS: Did this remotely and of course the connection dropped for each change - it's funny how much panicking you can fit into in a couple of seconds.

Hmm.. there is definitely something in this one i.e. I've noticed it happening on both 32-bit 2008 and 64-bit 2008 R2 servers, but not very often on 2003. This is different sites, but common features are: - They are all DCs. - They all have McAfee 9.7i Patch 3 or 4, but if that's the cause then it's a serious regresssion because I remember that messing with the net redirector and causing this a few years ago - they eventually fixed it. - Most have IPV6, SNP, new autotuning features and so on turned off. I'm not seriously worried but I would like them gone. Something I haven't gotten around to yet is actually looking at the event data ... it might not be that useful but there is an error code i.e. extra clue, somewhere in there.

What about here?: strError=createExamFile(centre[0], centre[1]); if (createExamFile(centre[0], centre[1]) == null) You're calling createExamFile() twice... probably need "if (strError == null)"

Every year. A bit of smoke & mirror going on and ditto for the police. For LAs that's a 25.5% cut phased over four years... that's just me quickly with a calculator, some folk who have apparently looked harder at what was said have it at 28%. Here the council have already said it needs to cut at least 1 in 10 (non-school workforce) jobs and a quarter of managers.

Take it you've all seen the Multipoint 2011 beta?

They're essentially the same. Without seeing the cert myself there are two potential solutions: 1) Get the site to give you a copy of the CA cert and install it (double click & follow instructions, unless you're logged on as admin trying to fix this for everyone who uses a shared machine, in which case put it in the *machine's* trusted CAs). 2) Ensure the latest MS CA update KB is installed - this only applies if the cert was issued by some particular CAs e.g. I think IPSCA & GoDaddy both had installed-with-IE root CA certs expire in the last couple of years.

I think the most effective way of sorting out menus is to redirect them so that they only get shortcuts you put there. Some folk redirect to server shares, but for various reasons I sync them to the local machine during computer startup ("them" = several different menus for different user groups). Note that only sorts out the Programs bit, you still need to fiddle with GPOs to control everything else.

It's apparently already been happening i.e. there are some stats somewhere (couldn't quickly find the link again sorry) that say public sector jobs decreased by ~40,000 in the first half of this year. Times that rate by the timescale and you're 3/4 of the way there[1] --- I did late lunch and watched it - the most interesting bit of course is that schools are approximately "safe" i.e. although there may be a bit of Peter & Paul going on re. the alleged budget increase, when you factor in pupil numbers and so on. The devil & the detail. Have resolved to encourage Sprogette to be a rocket scientist when she grows up. Only flaw with that plan is: a) I know a couple of science folk working for a household name who got made redundant recently, b) T'other half is a Creative[tm] so there will be a bit of a tug of war. [1] Plus, not sure whether this counts here, but apparently everyone working for *those* banks has been included in the "public sector" employees total and it was a massive number. Undo that at some future point and...

Ooops - misinterpreted what you were getting at there, sorry.

Service Management Portal -> Co-branding (select in menu on left) -> Windows Live Admin Centre (link at end of first paragraph in the middle). Yes. If you access it via a CNAME you will get a cert error i.e. the cert isn't for mail.what.ever.it.is. The only relevant CNAME I'm aware of is the random-ish one you make to prove you control the DNS for your domain.