Blog du Tristank : Windows Vista

Improve Hyper-V Performance With Standard VGA!

tristank — Wed, 05 Aug 2009 06:36:27 GMT

Yes, kids, if you’re finding that the Hyper-V performance ain’t what it used to be since installing that whizbang graphics card driver on your shiny new seven core hyperthread-and-a-halved megaturboserver thing, you might be suffering from flushes.

Read all about it here:

Video performance may decrease when a Windows Server 2008-based computer has the Hyper-V role enabled and an accelerated display adapter installed
http://support.microsoft.com/default.aspx/kb/961661

KB titles are getting ever more catchy, I think you’ll agree. So many words, and we still couldn’t find space for “is”. Ah well.

So, in short, get back to your standard VGA driver (just uninstalling the whizbang one is typically enough for that) and you’ll be sweet.

Old MPSReports

tristank — Tue, 12 May 2009 05:14:21 GMT

There’s a new MPS Reports version in town, with new features : new 64-bit friendliness, various forms of wizard-driven hotness for all the products the individual old tools used to support, etc, etc.

Call me old school if you want, but I typically prefer the convenience of “run this and send me the CAB file”, rather than “grab this, install the prerequisites, and choose the following options in the wizard, then send me the CAB file”. For newer OSs, that’s a non-issue as the pre-reqs (.Net 2.0 and Powershell) are built in; for older OSes, not so much.

A colleague sent me a set of direct download links to the old set, so I’m going to publish them here.

MPSRPT_Alliance.exe	MPSRPT_Alliance_Readme.txt
MPSRPT_Cluster.exe	MPSRPT_Cluster_Readme.txt
MPSRPT_DirSvc.exe	MPSRPT_DirSvc_Readme.txt
MPSRPT_Network.exe	MPSRPT_Network_Readme.txt
MPSRPT_SETUPPerf.exe	MPSRPT_SetupPerf_Readme.txt
MPSRPT_SUS.exe	MPSRPT_SUS_Readme.txt

For whatever reason, the download pages to these editions were removed when the new version was published; personally, I’d have suggested that the new was added alongside the old – the old, for all their limitations, are well-understood and widely used.

But the files are still there, at least for the time being.

Update: Looks like the PFE edition is still available in a not-through-the-back-door way (thanks, PFE, you rock! Hey, *I* work for that organization! Yay!), and it’s the core old-school goodness you’ve come to know and love from MPS Reporting.

http://www.microsoft.com/downloads/details.aspx?familyid=00AD0EAC-720F-4441-9EF6-EA9F657B5C2F&displaylang=en

.HDMP and .MDMP files

tristank — Tue, 23 Sep 2008 11:12:27 GMT

Just a quickie – the rule is blog what you know, but I figure my speculation might be good enough here.

A friend gave me an HDMP file and asked what I could make of it. After the usual “I could make a hat! Or a brooch! Or a dinosaur!” type stuff, I realized it wouldn’t open anyway.

In my experience, most .HDMPs come with matching .MDMP files. I think of these as Minidumps (in the “real” mini sense – just information about threads and thread stacks), and Heap dumps (everything else the process knew or cared about in User mode).

This HDMP wasn’t openable in the debugger directly, but if its corresponding MDMP was present in the same folder at the same time, I reckon it woulda.

The feared WER-wolf produces these files in pairs (that’s Windows Error Reporting, kids, don’t be too scared, except that it invalidates everything we used to know about AEDebug registry keys and similar, but that’s another story for another time), and that’s how I’ve analyzed them in the past. I remember hearing of some sort of merge operation that needed to happen between M and H dumps, but I’m reasonably certain I haven’t bothered with that (I assume I’m lazy by default), so I think the debugger just does it for ya.

Now I’ve written that, I’m going to go look for references to support my assertions!

949180 How to create a user-mode process dump file in Windows Server 2008
http://support.microsoft.com/default.aspx?scid=kb;EN-US;949180

(At the bottom – mini and heap dumps - yay me!). Think that’s enough for today. Hugs!

Mesh Gush

tristank — Wed, 20 Aug 2008 11:15:46 GMT

Hi Everybody! I've been laying low for a while, in read-only mode, sorting, filtering, evaluating and generally catching up on stuff!

So, why break radio silence now? Well, I'm popping up to offer a quick endorsement for the Mesh platform, which was recently opened up to Australian testers.

I'm excited. I love it, and it's not even finished.

Right now, you could liken it to Foldershare plus Skydrive plus Remote Desktop, just with the out-of-the-box stuff in the preview.

But there's more to come, and the glimpse you get is absolutely compelling.

I love that I'm a mouse-over and click away from Remote Desktop to any of my machines.

I love that I can synchronize folders between any number of my devices, and have a copy kept online in my Live Desktop.

I love that it's a platform, and having seen some demos of what's possible from a programming perspective, I can't wait to get my SDK on!

At present, there are some To Be Implemented features, the odd glitch and so on (as you'd expect from a CTP), so feel free to ignore me for a while, I'll just say "told you so" later! :)

One tip I'll pass on: Quite a few people requested a versioning feature for files and folders in the Mesh forums (and note that right now, folders you share with other people are writable by them) - Windows Vista has this built in already, in the form of the Previous Versions feature that you can use to recover a document in a given folder - just get the Properties, and then check the Previous Versions tab for the folder (or document).

Go get a Mesh!

That Memory Leak Revisited

tristank — Wed, 25 Jun 2008 08:59:22 GMT

While searching for memory leaking troubleshooting techniques that could be applied to 64-bit Windows (for the DHCP Server memory leak I found I had the other day), I stumbled across the answer to my problem in an internal tool (weird that I missed it from a web search the first time, but c'est la vie).

A Windows Server 2008-based DHCP server that is configured in a workgroup environment may consume too much memory

http://support.microsoft.com/default.aspx/kb/949530

And that's my problem! One REG command (and one restart of the DHCPServer service) later, I'm waiting to see how it went, but it all looks promising, based on that article. Neat-o.

Windows Server 2008 Diagnostics Off The Cuff

tristank — Fri, 20 Jun 2008 14:15:47 GMT

A word of caution to those of you that like endings: this isn't over yet.

I'm running a rather sad and noisy X64 desktop as a server at home. Once a proud warrior, actually, no, wait, it was never any good. It's just a Virtual Server host (it's not quite Hyper-V capable; next one will be). SBS 2003, an IIS and an ISA Server all exist(ed) happily in there at one point. (Did I mention I virtualized my work desktop machine the other day? So liberating!)

I blatted Windows Server 2008 onto it at RTM, and it's been happily puttering along doing the RRAS internet access and Virtual Server thing for me ever since.

Until Recently

But I've had to reset it from unresponsive-no-mouse-no-capslock situations on about four occasions over the last two weeks, and as the problem wasn't getting any better, so I figured I'd take a look at what I could do to try to diagnose it.

My guess was that I had a kernel-mode memory leak (a user mode memory leak shouldn't ever trash the box to that extent), but it didn't seem to correspond with any driver upgrades or software installations... something else had changed, sometime.

Perfmon (the new, shiny version) or more specifically the Reliability Monitor confirmed my suspicions:

(happy, everything-used-to-be-so-nice side on the left, then the gradual decline due to Disruptive Shutdowns towards the right). Note the quite-regular interval of red things on the bottom row. (Does it happen more when I'm at home, he wondered?)

Preparation:

As I had a theory in mind, I thought I'd create a Perfmon BLG (log file with lots of counters in it; lots of people seem to like CSV, but BLG is faster, and I'm never going to be opening it in Excel anyway).

How to do that? Things have changed: now, I create a "Data Collector Set", it seems. Oh yeah, reading manuals and/or following basic instruction: not my thing.

I created a new one based on the System Performance collector set, which matches my needs nicely because it contains all the Process counters and Memory counters. Between that lot, I should easily be able to spot a memory leak.

Started the collector set, and made a mental note to check in tonight.

Tonight:

After a little fiddling, I worked out that the animated "Data Collection In Progress" screen wasn't generating a report, and that I'd have to stop the data collector set to view it. Right on!

So, one stopped data collector set later, the Reports view is what I'm interested in.

Remember your training - you're interested in patterns that have slopes or steps. One counter leapt out at me, which I moused over and found was....

Process (_Total) Pool Nonpaged Bytes

So, yep, there's a memory leak, and it's in one or more of the objects tracked by Process counters. So let's add the Pool NonPaged Bytes counters for <All Instances> (so I can see all the processes).

So Add all them, and there's a counter that matches the slope, but at a different scale. Click it in the display to select it, and it's SVCHOST#10. Hide all the other counters I've just added (multi select, right click, hide all), and then right-click it and choose Scale Selected Counter.

Whop! Matches the curve almost exactly.

So, now I know it's a service host, but I don't know which one (they all look alike to me). I assume it's probably still running, too. How do I find that out now?

Easy: Add the "ID Process" counter for svchost#10 (#9 pictured, artistic license)

And then click the counter in the list to see the value it has (the plotted line is flat across the graph, meaning it didn't change at any point). I get PID 1348.

TASKLIST /SVC tells me everything I need to know (well, not everything obviously, but enough to take corrective action).

Yep - it's the DHCP Server instance of SVCHost that's apparently leaking NPP, a kernel resource.

Why!? And why now!?

The graph tells me the times at which this happened, but the Event Logs are very, very quiet around then. So I'll need to use tracing or logging or some other technique to actually track down the cause of the problem.

I right-clicked the SVCHOST instance with PID 1348 and chose Create Dump File (awesome feature, mentioned that before), for archival/root cause purposes - it may well not be possible to see the cause of the leak after the fact from a hangdump, but it's worth grabbing just in case - and then restarted the DHCP Server Service.

Taskman memory use dropped by about 100MB straight away. This is not a busy network, and NPP isn't typically used as cache by user mode programs (he giggled (in a manly way)). Something weird is going on there.

I restarted my performance logging, and I'll check in again tomorrow to see if there's any further indication of a memory leak (I haven't done anything to fix it, so I assume there will be). Now, time to look for logging and diagnostic options...

A word on Perfmon in Windows Vista and 2008: USE IT!

If you're doing any level of performance analysis of Perfmon logs, you need to try out the new, improved Perfmon in Vista. It runs rings around the old one. It's fantastic (at least by comparison). It's worth the cost of the upgrade alone. Seriously, if you do any sort of work with perfmon logs, try doing it on a Vista box and see whether your life is 1000% easier! I'm not saying it's perfect, but by comparison with the last version in XP/2003...

What's in IIS 7.0 for me?

tristank — Fri, 29 Feb 2008 06:16:00 GMT

While having a seemingly-innocuous chat with a colleague, I was asked to "throw together a few points" on what IIS 7.0 would do for a web application I've worked with in the past. Serves me right for talking to people, really.

In this application's case, authoring, publishing and content creation weren't as important as eventual scale-out and actual application performance (otherwise the just-released FTP7 would have been top of the list).

Here's what I came up with, off the cuff:

Reduced attack surface/patching requirements: IIS7 is now modular, and only the parts of the server actually needed by the application are required to be installed. This can also enable performance increases due to reduced memory footprint.
Vastly (imho) improved compression capabilities and compression performance (with automatic compression back-off when CPU use is high! How cool is that!?)
Simplified web farm management - if the application is able to work reliably in a web farm scenario, the IIS configuration store can be centralized and shared across all machines in the farm. (also see MSDeploy)
XCOPYable configuration - easily ensure settings are consistent between dev, qa and prod environments.
Failed Request Tracing (aka FREB) - if failures occur, request traces can be captured that include detailed diagnostic information.
Windows Server 2008 security, performance and scalability improvements - the most secure Windows yet (I expect); support for the latest hardware and 64-bit computing; and optimized TCP/IP performance.
SSL performance greatly increased - Kernel-mode SSL (not the best link for admin types) reduces context switching and ring transitions, and improves performance
Better UI! After a short time working out how the new model worked, I'm sold.

Windows Server 2008 Desktop Edition

tristank — Fri, 29 Feb 2008 05:44:00 GMT

WS08 Workstation, Anyone?

Apart from those pesky Live applications that won't install (why would anyone ever blog from a Server OS, as I'm doing now? Beats me...), I've had a relatively good run with Windows Server 2008 (x64, naturally) as my desktop OS.

Key drawbacks:

As mentioned, no official support of the Live suite (thanks!)

My Catalina handset is the only audio device I currently have a driver for, and Muse sounds odd through the speakerphone.

No Hyper-V, cos I'm only on a P4-class machine without VT/Pacifica/hardware support.

Now, onto less desktop-py things - managed debugging on the shiny new platform, with WinDBG:

Managed Debugging of dumps from, oh, say, a Sharepoint 2007 server on 32-bit

As a you'll-notice-this-when-it-hits-you mention, the version of MSCORDACWKS that ships with WS2008 looks like it's 2.0.50727.1434 (unless that came from somewhere else) instead of 1433 (2.0SP1, if I recall correctly (and I often do)) - meaning that debugging a managed dump won't work straight out:

PDB symbol for mscorwks.dll not loaded
Failed to load data access DLL, 0x80004005
Verify that 1) you have a recent build of the debugger (6.2.14 or newer)
            2) the file mscordacwks.dll that matches your version of mscorwks.dll is
                in the version directory
            3) or, if you are debugging a dump file, verify that the file
                mscordacwks_<arch>_<arch>_<version>.dll is on your symbol path.
            4) you are debugging on the same architecture as the dump file.
                For example, an IA64 dump file must be debugged on an IA64
                machine.

You can also run the debugger command .cordll to control the debugger's
load of mscordacwks.dll. .cordll -ve -u -l will do a verbose reload.
If that succeeds, the SOS command should work on retry.

If you are debugging a minidump, you need to make sure that your executable
path is pointing to mscorwks.dll as well.

After a little fiddling with .cordll, I worked out the easiest method (read: the first method I hit upon that worked) to get past it.

In less debugger-y speak:

grab a copy of the 2.0 SP1 version of MSCORDACWKS.DLL from the Windows\Microsoft.Net\Framework\2.0.50727 folder on a machine with the right version (I fished one out of my Vista installation on the same machine),
paste and rename it into the same folder as WinDBG, as mscordacwks_x86_x86_2.0.50727.1433.dll (rinse and repeat with amd64 for the 64 bit version, from Framework64).

Then .cordll -ve -u -l should work, and so should your SOS commands.

Anthology Of Interest

tristank — Mon, 17 Dec 2007 17:24:00 GMT

Okay, so there's no interest here, but perhaps an anthology anyway.

Yes I've been gone a while; another computer packed it in (this time without my own personal brand of assistance), I've been off to India (usually 5.5 hrs difference from Sydney, but we worked the night shift - 10pm to 4:30am in theory, we made it to 3 one night and faded otherwise - for a few days there before rotating back into pseudo-reasonableness of 1pm to 10pm IST, and have been jetlagged for the two weeks since returning... I know, a tough life).

Anyway, quick mentions:

Touch Ain't Just For Tablets

Mary Jo seems to think Touch is for Tablets only... Not quite. Perhaps Microsoft agrees, but why is Touch considered a Tablet feature only? It's brilliant anywhere, if you can get it.

Is it really that expensive to add to a regular laptop? Shouldn't it be a standard feature?

After having the p1610 for a year now (they grow up so fast!), I'm thinking that the tablet form-factor novely has utterly worn off (it only ever gets converted by eager demo-seekers, I use it in clamshell style only).

So, upgrading and replacing it might be on the cards sometime - an ultralight, say a 10.4" to 12.1" wide screen would be fine. Small keyboard is fine, I've adapted perfectly to the 1610's 3/4 size keys. Give me long battery life and light weight, and I'm happy, mostly.

But, mandatorily, any future laptop I buy must have a touch screen.

I don't care if it's landscapo-portraito-tableto-converto-capable, or if it does handwriting recognition (I have issues recognizing my handwriting; how's a computer ever going to get it?), or any of the other (actually very cool) Tablet-ty features; it just needs that if-you-push-a-spot-on-the-screen-something-happens capability.

I made my last purchase decision based on the device being one of the few that had a touch interface (as opposed to a Special Pen interface, which not only makes the pen sound challenged, it's a challenge to fish the pen out and cause of repeated heart-stopping "I forgot the pen at the client site!" moments, before you discover it's in your other pocket, or caught up in your belt, or whatever) and it was, in retrospect, a winner.

Both long-time readers will know that I don't put a lot of faith in my gut feelings (without lots of testing), but this one panned out brilliantly.

Jabbing at the screen with one finger is brilliant.

As an LCD purist that actually puts a "THIS IS A FLAT SCREEN NOT A TOUCH SCREEN - FINGERS MAY BE LOST" label on his precious (fingerprintless) LCD monitors, I can honestly say I'd convert every single one to a resistive touch screen covered in finger goop, if it got me proddability. It's the very definition of intuitive. You see a button, and you push it, physically, no mouse-hand-translation, no infinite widths or screen edges required.

Your finger is the original pointing device. (Er, arguably. Let's not go there.)

If every laptop came with a touch screen (not an active digitizer, remember), just imagine how much less frustrating plane trips could be? Instead of scratching like a chicken at the pad, or nudging nervously at the "pointing stick", you'd poke your way through dialogs and drag your scrollbars down directly.

Happy sigh. So as you can probably guess, I'm very interested in any laptop sporting a touch screen (active digitizer too for bonus marks, sure, but it's primarily about the convenience - having 1024 pressure levels is a secondary concern to me, most of the time). But I want my fingertip (or pen, or ice cream stick, or knuckle, or tightly-coiled tissue) interface first.

Laptop makers: you know it makes sense! Make it so!

You Know It Shipped When...

Visual Studio 2008 shipped a while back now (er, belated yay!) and installs with the .Net Framework 3.5, which includes 2.0 SP1 and 3.0 SP1, in case you didn't know.

So: installing 3.5 doesn't just do a bolt-on installation, it does actually upgrade the core 2.0 and 3.0 binaries. The SP releases are available individually too.

Have a great holiday season!

Tip o' the Week: WEVTUTIL for EVTX/EVT file conversion

tristank — Fri, 05 Oct 2007 09:19:19 GMT

This week, a pointer to a solution to a problem I occasionally hit.

Windows Vista (and by extension Windows Server 2008, I assume) utilizes a new EVTX log format for event log exports. It's XML-based, natch.

Problem: Everyone's Favourite Log Digestion Tool Log Parser uses system APIs to read event log exports, and the old .EVT event log format isn't "native" any more. Long story short, it chokes on them.

This, to put it mildly, was annoying, as most customers haven't moved to Windows Server 2008 yet (I mean, it's only five months from release - is there ever a better time?) and so supply event logs in the old format when asked.

Anyway - you can convert the old-school event logs into shiny new event logs through the user interface (just double-click the EVT, wait for it to open and display in chronological order; then do a Save As, pick a location and filename and answer an obscure question about language formatting; then find and open the newly-resaved log file), but bluntly, the GUI process leaves a bit to be desired if you have the slightest inkling towards type-A behaviour, and all I really want is something that'll work in Log Parser, really.

WEVTUTIL (and NeilCar) to the rescue. It's included out of the box, and it'll convert those dusty old event logs from the command line, with nary a GUI or common dialog in sight, ready for consumption by Logparser, or any other EVTX-friendly file muncher.

Neil's example (for the click-inhibited):

wevtutil epl application.evt application.evtx /lf:true

Bewdiful.

Netmon vs Chimney

tristank — Mon, 23 Jul 2007 06:22:00 GMT

I recently encountered TCP Chimney for the first time in the wild.

Short version: Chimney is an offload technology that allows the NIC to deal with up to X TCP connections, with any overflow being handled by Windows. All good: get the NIC dealing with more networky stuff, and reduce CPU use. Excellent!

The reason it came up:

I was staring at a small network monitor capture (should have been much bigger) that should have had a few tens of megabytes of FTP but was mostly comprised of SYN, SYN-ACK, ACK s to port 21.

A lot! It'd look like

SYN -> 21, source port X
SYN-ACK
ACK
SYN -> 21, source port X+1
SYN-ACK
ACK

...Hundreds and hundreds of TCP 3-way handshakes, but next to no actual data sent. The server didn't even appear to be sending its connection banner!

Very, very rarely, I'd actually see a frame or two of FTP traffic, but I thought the symptom I was looking at was indicative of resource starvation on the FTP server.

Perfmon didn't confirm the diagnosis, and the FTP server logs showed it was transferring loads and loads of data; I just couldn't see it in the capture.

After ruling out network adapter teaming (The Old Enemy), I wondered if something from the Scalable Networking Pack might have been involved, and a quick internal search later, whop! A symptom match! Because the NIC handles the heavy lifting of all TCP work with Chimney enabled, after the TCP session is established, Netmon doesn't get to see the traffic!

To disable Chimney so you're able to gather captures for troubleshooting purposes, you can use the following netsh command:

Netsh int ip set chimney DISABLED

Once that's done, Netmon (and presumably other NDIS capture drivers, like WinPCap (ethereal/wireshark) should be able to capture all traffic, not just non-TCP stuff!

Fear The Hand Of Non-Selection

tristank — Sun, 22 Jul 2007 09:30:53 GMT

One of those "remember it for when you need it" posts:

On Tiny, I've been unable to select text in the Outlook preview pane for about a month. When I tried to select things, I'd get a not-quite-looking-like-the-web-hand hand instead of a selection bar, and it would grab at the document rather than just sorta sitting there browser style.

It hasn't been a big deal, but I couldn't find anywhere to turn it off in the options, or work out what was happening.

I wondered if it might have been new-fangled Outlook behaviour on Tablet PCs, but couldn't reproduce it on any others.

Cutting a long story short: I found the culprit - it kinda was a tablet thing, but more due to my use of the scroll bar on a tablet.

At the top of Word scroll bars (and I guess this flows forward into Outlook 2007, which uses Word for email rendering), just above the top scroll button, there's a little hand.

Clicking that little hand (intentionally) puts you into drag-the-document-but-don't-modify-or-select-anything mode.

Clicking that little hand (unintentionally, and less likely to happen if you have a mouse wheel rather than a tablet) will cause the same effect, only you won't be able to work out what happened or how to undo it! :)

I happily accidentally ended up in scrolly mode today while reviewing a Word document, but noticed the effect and was still staring at the button I'd failed to tap.

Yay!

Two New Updates for Aussie Windows Vista Media Centre Owners

tristank — Fri, 15 Jun 2007 08:11:00 GMT

The Media Center dev team have been working to address a couple of problems that have particularly affected some Australian Media Centre users, and we now have updates available from Product Support.

While the fixes are available, the KB articles documenting them aren't yet published (at the time of writing), but we're working on getting them out as soon as we can.

In Australia, you can call Support on 13 20 58 (1,4) or use a pre-existing support arrangement (Partner, Premier, etc) to obtain either hotfix directly from Microsoft PSS.

If there's a problem getting them sent to you, you may need to explain that the articles are still in production, but the fixes are ready and available from the hotfix site.

As with all hotfixes, our recommendation is that if you aren't affected by the problem directly, you should wait for a future public update or service pack containing the updates.

Media Center is available in the Windows Vista Home Premium and Ultimate editions.

First, there's an update that provides a more reliable way of keeping custom Guide data in the registry, without having to adjust windows services (for example, KeepKey), registry scripts, scheduled tasks or registry permissions.

935685 - MCUpdate Crash when disableUpdateDiscSvc is enabled * - KB article being finalized. When published: http://support.microsoft.com/kb/935685

In short: the update allows Guide providers to set the disableUpdateDiscSvc registry value to 1 without crashing the MCUpdate process when it runs, to retain their custom guide data registry key.

I would expect that you'll likely be directed to install the hotfix by your EPG provider, and they will set the registry key with their installation/uninstallation, but check with them!

In case technical details are wanted:

Registry location: HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Media Center\Service\EPG
Value: disableUpdateDiscSvc : DWORD : 1

Removing the value or setting to zero should return to normal behaviour (the discsvc key will be overwritten when MCUpdate runs, returning the guide to its default setting).

This next problem seemed to crop up fairly recently on my installation: when channels retitled themselves (as often happened with Channel 7 during a special event or similar), MCE (or rather VMC as folk are calling it now) would sometimes assume it had found a new replacement channel, and forget about the old channel on the same frequency. This led to a loss of Guide data on an occasional basis (and scheduled recordings would often not work on the new channel), It affects Terrestrial Digital TV stations only, far as I know.

938927 - Opportunistic Scanning causes loss of EPG/Guide data * - KB article being written. When published: http://support.microsoft.com/kb/938927

Once the hotfix is installed, opportunistic scanning can be disabled by setting a specific registry value:

Registry location: HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\ CurrentVersion\Media Center\Service\GLID
Value: DisableActualStreamOpportunisticScanning : DWORD : 1

This one you'll need to install and enable yourself (unless the guide provider does the registry key for you - they're generally very nice people, so expect to see a tickbox in their programs in the future).

We've had good results in testing, but you might want to consider rerunning Media Center setup after installing the hotfix to rescan all your channels from scratch - if the channels are already weird when you install and enable, they might stay that way.

Again - neither KB is done yet, but we're working hard on them. In the meantime, contact Support to get the updates directly from us.

See also Mike Hayton's post in the XPMediaCentre.com.au forums.

This post is provided "AS IS" and confers no warranty.

IIS 7 - LH Beta 3 Launch Stuff

tristank — Thu, 26 Apr 2007 11:45:26 GMT

Tons of cool new things in Longhorn Beta 3 on the IIS front, including:

... the first beta of our new FTP server. This isn't a part of Longhorn Server Beta 3, but it is available today and offers a lot of brand new features for you to experience today. The new IIS7 FTP server includes secure publishing with FTP/SSL support, integrated web publishing with support for the IIS7 configuration system and administration tool - making it really easy to setup FTP publishing points for a web application, integrated authentication (so that delegated administrators can use their same credentials to publish to IIS7 web sites), host header FTP support, and more! This FTP server is really amazing, and I'll be covering more of it in a future blog post. Download it today, and read more about it on IIS.net!

There's also a new DevCenter (or Developer Centre for us posh English-speaking types) with an expected-to-be-ever-expanding selection of samples and tools.

Goings On...

tristank — Mon, 19 Feb 2007 08:49:11 GMT

Quick catchups:

While in Seattle, I talked with some folk about IAG and ISA 2006 - all I can say is *wow*. I'm downloading the demo VMs now to have a play.

The other *wow* was SoftGrid. Architecturally, wow! I'm eager to learn more about that.

While on the *wow* subject, I upgraded my parents' PC to Windows Vista. This is a parts-up-to-5-years-old Athlon 2100+, with 512MB memory and a 200GB hard disk (relatively recent). At first, it was slow. I mean, super-good-grief-will-it-ever-open-Outlook slow. It didn't seem to be memory, as there was still a bunch cached, but I thought I'd try some more anyway and jammed in a gig - no help (see, this is what you get for troubleshooting based on opinion rather than evicence...). Then, we realized that a) the antivirus software was being a PITA, and b) the networks were reversed - the internal network card was considered public, and the external one private. I'd already disabled the antivirus by the time I noticed that, but it was then that the box calmed down and started performing really well. So I left the extra memory in anyway, and I'll surprise them with Office 2007 when I'm next over there.

Is there a moral? Nope, just a data point.

Actually, while I was over there for something else, I heard a really good description from someone explaining to a non-computer-person the difference between Windows Vista and Windows XP: "You know your 2001 BMW? This is like a 2006 BMW. Same thing, 5 years better."