Microsoft gets personal on Windows 7 "show stopper" bug Microsoft has gotten personal in responding to reports of a "show stopper" bug in Windows 7 capable of delaying the planned roll-out, which starts Thursday. The company has blamed a chip-set controller issue rather than a critical bug in the Windows 7 chkdsk /r tool that could cause a memory leak capable of causing your PC to …
Reminds me of the BS that Microsoft tried to peddle when Windows XP started locking up USB 2.0 ports on some motherboards.
They blamed the chipset firmware, yet the Windows 2000 drivers worked perfectly with the same firmware, so the BIOS firmware providers not unreasonably pointed the finger right back at Microsoft, saying they should fix XP since it was clearly XP that was broken, not the firmware.
Stuck in the middle of this finger pointing exercise is the punter. i.e. me!
In that case there was a painful workaround that involved duping XP into using the Windows 2000 drivers, but in the end I simply disabled my on-board USB ports and installed a PCI USB card, which gave me more ports than I had on the mobo anyway.
Paris, because surely it's de-rigeur for any comment referencing "firmware"?
"Show stopper"? What's the first thing to do after installing Windows? Is it:
(a) run Microsoft Update to apply the latest patches; or
(b) run ChkDsk?
As for Microsoft pulling the RTM - does anyone seriously expect there won't be a new bug in IE8 discovered before the end of October? This is simply business as usual - not just for Microsoft, but for any large software business.
That this turns out to be a driver issue. I can't really see how a 'critical' bug of such magnitude would have suddenly appeared in Win7, because as we all know, Win7 is really Vista service pack 3 R2.
Right, now that we've had a level-headed thought on the matter, time to roll out the usual tidal wave of uninformed ignorance.
<rings bell>
BRING OUT YER TROLLS!
WHAT testing process?
The really funny part is when the MS shills claim that's not a flaw it's a feature.
Sure, it's a feature that your PC run out of memory. Didn't MS ever think to put limits on how much memory it's using? Seems like one of the first critical design parameters when they decided to *fix* the formerly working chkdsk.
Seems like Vista had a similarly incredible flaw where you couldn't even copy a few thousand small files without it taking days. Seems like Home Server wouldn't let you use one of the only real reasons to that can't be so easily bolted onto vanilla XP, the virtualized filesystem.
Wasn't there a statement that MS hadn't reproduced the problem on their 40 PCs? They have a mere 40 PCs, and yet none with the Intel chipset they try to blame? Oh wait, it's not even likely that it's isolated to one chipset, just one manufacturer's chipset that happens to be the most popular world-wide manufacturer of PC chipsets?
MS you really are lost. The average PC owner is not exposed to your PR damage control, they are not impressed when something is fundamentally broken right out of the box regardless of where the finger of blame is pointed. If this were an obscure bug or you were a poor startup, 2 guys working out of their garage I could see it. Someone is bound to write something like blah blah blah googlezillions of lines of code and every OS has bugs. To that I answer, if you can't bugcheck the new code then leave it alone, but this was not a flaw in programming, it was a design flaw in the way the revised chkdsk works.
Driver or no driver there should never be an unbounded limit on how much memory it can consume.
"Bugs that are so severe as to require immediate patches and attention would have to have no workarounds ..."
OK, so that's the status of their "trustworthy computing" initiative. If there's any kind of workaround available, the flaw doesn't deserve any immediate attention.
I think that explains a lot. Any questions?
Hes right, its is very unlikely MS would delay the release of W7 because of this ...
Since when has a bug that locks your PC up and causes it to crash prevented MS from releasing stuff?
OK, its minor, im still annoyed at what seem to be their plan .. "well, you insisted on sticking to XP because Vista was a pile of crap .. so, we've done what you asked and built a totally new OS for you, that is nto Vista SP3 no sireee, it all new is windows7 .... err, please don't look under the hood, thanks".
it just another attempt to railroad us into what is Vista by another name, so they desrve as much stick as they get :)
"chkdsk /f is run automatically on detecting a volume that needs checking?"
That happens only rarely for NTFS volumes. In most cases the journal entries are just rolled back to a safe state. It can still happen (and I had to recreate a VM once because ChkDsk became stuck) but it's unusual.
If anyone is storing important data on a FAT volume they need their bumps feeling :)
Paris because she likes her bumps feeling.
I've got linux on a few boxes at home and use it at work, so certainly not an anti-linux type, but a whole bunch of untested bug fixes hacked together in some sort of 'my cock is bigger than yours because I can get a fix made first' competition is not a fix to a problem. There is the slight matter of testing and waiting for the selected fix to be put into the stable releases. Ultimately pretty much the same thing would happen with a new FOSS OS as would in Windows or any other commercial OS, if a bug is identified post release - a fix will be made, tested and released.
Well, the thing is, if they took such little care of this item, how little care did they take of other important things? And don't mistake a memory leak for "insignificant".. if you can get a machine to keep committing memory even though it doesn't need it, there's an open hole waiting to be exploited.
Paris... because she also likes being exploited
This is not the first serious defect to crop up in the RC code; I reported a bug in 7100 which would cause all users ona Win 7 system (including administrators) to completely lose permissions and ownership of the main system drive... This would happen quite randomly, usually when rying to make GUI changes to network settings.
I was under absolutely no doubt whatsoever that this was a 'critical' defect and although I was criticised for reporting it as such in the Win 7 forums, soon after there was an absolute flurry of new releases during which the bug appeares to have been fixes as I haven't experienced it since I started using build 7232 and more recently 7600.
No it doesn't. I've used Win 7 Beta 3 and RC1 for the past, ooh, 4 months at least for day to day use and not seen that bug. It's faster and lighter than Vista, much better engineered and a lot more enjoyable to use. My OS of choice is Linux and my desktop of choice is KDE4 but I'm actually very happy to use Win 7 at work, it's much much improved and they've got UAC toned down to sensible amount of prompts...
If Microsoft haven't reproduced the crash or experienced any crashes with chkdsk on the stack reported in "any measurable number", how do they know that updating the chipset drivers fixes the problem?
For at least 20 years, Harddisks have hidden bad sectors from the OS by automatically mapping pre-reserved good sectors into their addresses. A fault in chkdsk /r itself is not something many people will ever bump into, but if chkdsk is exposing a bug somewhere else that other programs may trigger then there is a problem.
Clearly, the best way to solve this type of problem is to post a random 'solution' before getting clear results from a proper investigation, and to blame the people reporting the bug.
Yes, it does. Win7 has the slow file copy bug, or at least the RC does. I have also been using Win7 at home for the past month or two. Tried copying 4GB of files between two 15000rpm drives. That took an hour. AN HOUR! I took a screenshot, if you'd like to see it..
Also tried copying a 300MB programming project folder from a USB stick this morning. That took 15 minutes. This is taking the piss. And while the file transfer window is up the rest of the system slows down considerably. And this is *not* a slow machine - quad core with 4GB+ of RAM and SCSI hard drives.
As for "nearly as fast as XP" - why the hell isn't it *faster* than XP? We're only talking about bloody file copying here - one of the most basic operations of any OS. Why should that ever get slower?
I like Win7, I really do. It's a much better upgrade from XP than Vista was, and I've gone so far as to pre-order it. But that bug is a real pain in the arse, especially when you're a programmer and folders full of code files are your thing...
Just spotted it somewhere else. Ran a Maven build script set to package a project into a WAR file. Normally takes around 15-20 seconds to assemble the project prior to packaging (on a much slower machine, running Ubuntu). On here it just took 197 seconds - over three smegging minutes!
Perhaps you might only run one antivirus instead of lots. Maybe you not good at protecting the computer you have so it has many spyware and porn linking on the desktop? If he try a clean build first then might be better before to criticise something that we know is not the case. Also we see that you use the Maven WAR files - this is Javs.
First of all, I have yet to see a single instance of anyone that claimed this crashed, or BSOD'd their system. If so, Send in the Crash dump file to MS. Second, This blame it on the chipset , or board manufacturer's in pure BS. If you want to know the truth, the issue was with chipset controllers for the hdd's. nothing more. The memory usage is design to complete the task faster for the user. If your running Chkdsk /r comand, I wouldn't be multitasking it with games and so forth. it's suppose to cap out a 90% if memory usage, leaving you with roughly 50mbs of memory residual while the tool does it's job. The only reason to ever run this tool is if you have a possible hard dravie failure or corruption. The is not something that you will use everyday, unless of course your some computer repair person. Either way MS will fix this before the public release. So , no worries! It'll get done. =)
The real problem here is that this bug is effectively unpatchable for people who use their Windows 7 install disc to boot in order to perform repair operations, a scenario in which it is entirely possible to want to run chkdsk /r.
Basically, we have a situation in which the standard recovery environment is unsafe even for those who've read all the manuals, since chkdsk is not generally regarded as a dangerous operation which needs special drivers (as long as you don't yank the power cord halfway through).
The RTM is done, so MS needs to fix this and provide a way for people to create a new, patched repair disc.
worried about releasing a product with bugs or "features". Let's face it. The people that buy M$ software are the real beta testers, not the bklokes cowering in the back rooms of Redmond.
If this was not the case then M$ would not need to continuously bring out patches for their OS and "productivity" products.
However, calling this one a show stopper is rather like George Shrub calling "Job Finished" in Iraq when he did, namely way off the mark.
Don't be a fecking idiot. You acknowledge that I'm a programmer (ie, someone with a modicum of computing knowledge, numbnuts) then suggest I can't keep the system free of malware that slows down file copying in exactly the same way as a bug you claim no longer exists?
And as for Maven being the cause of the slowdown, you obviously can't read. That's only one case in which it occurs, and it's much, much, much faster on Ubuntu. XP too.
Prat.
"Since when has Microsoft worried about releasing a product with bugs or "features". Let's face it. The people that buy M$ software are the real beta testers, not the bklokes cowering in the back rooms of Redmond.
If this was not the case then M$ would not need to continuously bring out patches for their OS and "productivity" products."
Of course they would. The chances of you writing a perfect operating system in which no bugs are ever found is very very slim, especially when your OS is 95% of the market (or whatever the current stats are). I wonder what percentage of systems have never needed to be patched after release. Actually, I wonder if that's ever happened....
What I have a problem with is the way in which MS prioritise their work. Critical security updates fall by the wayside, but you can bet your ass that the latest version of Genuine Advantage is out on time,
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
