Windows 7 customers hit by service pack 1 install 'fatal error' flaws A brace of "fatal errors" is hampering Windows 7-based computers that have been updated with Microsoft's first service pack for its current operating system. In fact, since Windows 7 SP1 was released late last month, many users have been grumbling on forums about problems with the install of the update package. Similarly, The …
...and various BSDs before that, I'd like to point out that all operating systems have bugs, even Linux. In the case of Linux, if it doesn't work it's not necessarily the fault of the developer. It can be the hardware manufacturer or the user that is to blame, but this hardly makes Ubuntu the answer to all software problems. If a user is left powerless by a service pack update, Ubuntu is just going to give them harder problems to solve.
I'm a big fan of choice and having the right tool for the job. Sometimes, and for some people, Windows is the right tool.
Simon
With Win 7, even without access to the original media, you can burn a recovery DVD from the backup control panel that allows you to boot and diagnose just the same as having the real OS DVD.
Also, failure to have enabled backups, snapshots, etc to be able to recover from a service pack failure is no excuse. Anyone "rebuilding" due to a package failure is simply doing it wrong, unless the disk itself got currupt of failed in the process (unlikely coincidence not typically related to the patching process itself, but a bad disk to start with).
"but at least if you're running a *nix system you can usually boot another instance and hack the config to sort it out."
This is exactly why so few home users have Linux installed. "Hack the config"??? What ordinary user knows what on earth that means? MS screwed up on this one, but don't tell me that hacking the config is an option for 99.99% of users....
This post has been deleted by its author
"If any piece of software has a bug it's the fault of the developers. The end. You can't push the blame on to hardware manufacturers because it wasn't tested sufficiently, or on users because they're allowed to do things they shouldn't"
Utter bollocks. Any compatibility and stability problems endemic to Linux are overwhelmingly caused by manufacturers not releasing proper (or any!) specs for the hardware, and kernel devs having to take a reverse-engineered "best guess".
Most device drivers for windows are written in-house by the manufacturers themselves and simply cross-signed with a Microsoft certificate. Free software seldom has that luxury.
What if the cause turned out to be a popular anti-virus software? How would that be Microsoft's fault?
It's not beyond the realms of possibilities that a 3rd party product could get in the way. That would suggest a lack of real world testing.
It might be interesting to hear of the numbers, a vocal thousand users is nothing out of a user base of millions.
I had a cheap consumer ASUS (?) multimedia mobo where Ubuntu would just hang at boot. Turns out the problem was caused by a crappy BIOS, probably involving power management/hibernate. Upon flashing the BIOS, the same exact Ubuntu worked just fine.
If a faulty BIOS is not the fault of the hardware / BIOS vendor, whose fault is it, exactly?
p.s. I've been using Linux on and off for about 10 years. Currently using OSX instead, because I really don't care much for having to do admin work on my primary machine. I just want it to work and am willing to lose some freedom in doing so. So it's not like I am a rabid Linux fan.
Yes, the hardware is a nightmare. I just gave away a new Toshiba which had awful problems. It seemed that there were differences from previous units with the same model number. Really, I couldn't be bothered, and bought a reconditioned Dell, which works ok... except for a microcode update on the wireless (maybe that's why it got sold).
It's a jungle out there.
I agree that it's a horses for courses market out there, and for some people, Windows is the Right OS.
Sadly though, Windows being the proprietary soup that it is, it makes it damn near impossible for the end user to debug.
I don't envy Microsoft's position here. They've got literally thousands of reports, all with various device drivers and user applications, many of which were developed in secret, most of which were developed in isolation. All of which can be interacting in strange and unwanted ways.
Yes, the hardware abstraction layer is designed to address some of these needs, but it still makes it one big mess for them to maintain. It's like the numerous medical specialists prescribing various drugs to a patient, leaving them with a cocktail of pills no one is sure about. I'm not surprised that things break.
I guess Microsoft have more power to coerce co-operation with the companies involved, but I dare say it does sap a lot of time and resources simply because the industry are used to behaving like proverbial cats.
The open-source solution to the problem isn't perfect either, and some companies run a mile from it, but I think the more open and collaborative approach is helping in this regard. I haven't had a breakage on a Linux machine that has seen me having to completely reload in a very long time.
Windows on the other hand, I've had installations spontaneously develop faults that have taken days to rectify.
Linux costs me my time to maintain it. Windows seems to cost both my time, and significant money, thus for me it's an easy choice.
I think that Microsoft could do well to look at how things like device drivers and such are managed in BSD and Linux… which in doing so, could make their lives easier, and improve the user experience for everyone. That doesn't mean open-sourcing the Windows OS (although that could have some good benefits), just co-operating more closely with the manufacturers and software companies to better co-ordinate efforts.
Re : sgb
Yes all os's have issues sometimes with updates, I have also seen Ubuntu installs no longer able to go into X or even get to grub after an Ubuntu (although the same applies to all distros) update.
The big difference is ..
Linux is always so much easier to fix the issue and get back to your data when you have an unbootable OS - mainly thanks to live cd's - normally you get fix the issue yourself rather than be at the mercy of some joke company to wait for a fix.
Windows 7 recovery process actually works now. I've had great success with the Windows media in recovery mode, or the restore disk made via the backup app being able to automatically repair windows installations. Its fixed things that under Vista or XP would have required multiple 3rd party tools or a complete restore. Not perfect (and still required FAR more often than reasonable), but it actually works. Comparing that process to Linux recovery, and the knowledge required, toss up on complexity and effort... Apple still has the best recovery process...
> If a user is left powerless by a service pack update, Ubuntu is just going to give them harder problems to solve, sgb
In a working environment, you don't ever-ever install someone elses software patch. If you do patch, it's because something isn't working, and then you only patch that bit and nothing else. If it ain't broke, don't fix it. And calling something a service pack still don't make it a bug fix.
> I'm a big fan of choice and having the right tool for the job. Sometimes, and for some people, Windows is the right tool.
Insert marketing slogan for the Microsoft organization. Personally I am always pleasently suprised when my Windows box sucessfully boots up each morning.
I've had perfectly ordinary Ubuntu installations self destruct after installing perfectly ordinary automatic updates. Took some serious hacking to sort out, and was way beyond what an ordinary user could be expected to do. Not very impressed.
MS have had, in my personal experience, fewer update problems.
Considering how important the ability to automatically update is to the adoption rates of an operating system, it's amazing how badly they go. An OS with a reliable update mechanism will gain a reputation for continued improvement, and things can only get better. An OS with a reputation for not getting successfully updated will be deeply unpopular because users know they'll get left high and dry. This never used to be a problem in consumer land - bugs meant crashes that customers just fixed by turning off/on. But as functionality increases the power cycle fix becomes less acceptable.
MS will get this one right sooner or later; they have generally done so in the past for mainstream windows. MS also have to get WP7 updates right too. Apple are just b*****ds because they use updates as a mechanism to piss off customers with older kit by leaving bits out for no sound technical reason, and even then they don't always get them right. Android is an update joke.
> I've had perfectly ordinary Ubuntu installations self destruct after installing perfectly ordinary automatic updates. Took some serious hacking to sort out, and was way beyond what an ordinary user could be expected to do. Not very impressed.
What were you doing running 'automatic updates', why didn't you restore from the nightly backups. You do keep nightly backups don't you?
> Considering how important the ability to automatically update is to the adoption rates of an operating system, it's amazing how badly they go. An OS with a reliable update mechanism will gain a reputation for continued improvement, and things can only get better.
See my other post, you don't ever-ever automatically update a working system. If it ain't broke don't fix it.
<raises hand> I've had problems too although thankfully not as bad as those cited in the article.
I'm running Win 7 Ultimate (non-OEM). I waited a couple of weeks after release before I installed SP1 just to see if the zero day adopters hit any major problems. Having not seen any big problems reported (not an exhaustive search for problems I admit!) I went ahead and installed only to find that Hibernate no longer worked on my desktop machine. If Hibernate was selected the machine shut off almost immediately, rather than taking 30-60 seconds to save the current state to disk first.
Luckily I was able to wind back to the restore point created when SP1 was installed, and the problem was solved. The commentard cited in the article is right - SP1 does seem to remove all earlier restore points when it installs. Well, at least my experience concurs with theirs.
Big fail (Again) M$FT.
>Surely Wsus-ers could just NOT approve the update? Who in their right mind lets all updates >auto approve? Especially a service pack!
Unfortunately something seems to have overridden the controls that allow you to only automatically approve certain updates. Possibly an SBS2008 issue...
----
This fixed the issue for us:
Boot to repair your computer, cancel repair, select Advanced Options and run a command prompt:
just do notepad exe, open reboot.xml
<?xml version='1.0' encoding='utf-8'?>
<PendingTransaction Version="3.1">
<POQ>
<DeleteKeyValue path="\Registry\Machine\COMPONENTS" name="PendingRequired"/>
</POQ>
</PendingTransaction>
copy that out and drop it into your pending.xml at the top
reboot and voila!
I mean, going by the Pratchett rule what you're telling me is that at least 4 people have had this problem.
How about some numbers on how many people/organisations using WSUS did testing before pushing out the fix, and can prove that the test machines were fine with the update while the production machines weren't? Are there "many" of those as well?
...comes from a joke in the Discworld novel "Men At Arms" about dumb trolls being unable to count because they would go, "one, two, three, many". The joke is that no one ever considered that "many" can simply mean something else in a base-four counting system, with each "many" adding four to the number (and with "lots" to indicate a many of many--16, by our standards).
I was an early adopter of SP1 (a couple of days after its release) and I enjoyed the following seamless experience:
1) My work Laptop. SP1 successfully installed via Windows Update after two tries,
2) My fathers Desktop. SP installation failed repeatedly but the system rolled back every time. Mind you, I didn't hit the power button like some end-users might do.
In the end, and after many failed attempts to defeat the evil SP, I took out the Windows 7 upgrade CD (which was fortunately available) and performed an in-situ upgrade of the OS, fully expecting the worst to happen. To my utter amazement, Windows 7 happily upgraded itself (although all subsequent WSUS updates obviously had to be re-applied) and then very happily swallowed SP1 after A SINGLE try. That method was recommended on an MS Blog in response to many desperate inquiries by people like meself.
I am not saying this was fun. And I'm dead certain I'm not the only one who was bitten in the ass by SP1. I am just happy I discovered the good news before approving SP1 distribution on our corporate network's WSUS server (but no one should do that with a major service pack anyway, and yes you don't have to approve every update with WSUS, but you do have to learn how to drive WSUS). So I wouldn't rush to roll out SP1a , I'd wait until MS fixes itsgremlins. Alternatively you can test it in a quiet corner on an unimportant machine if you have an afternoon to spare . And dats da truth, o skeptical ones. The sky does sometimes fall.
Ran into this problem with a Win 7 Ultimate install - actually restored from image - tried again. Boom. Did a swarch, and some obscure site mentioned that the errors seemed to be something about memory being exhausted.
Apparently - if I follow this correctly - MS decided that instead of patching the language packs, it would have the SP remove them. For some reason (trying to do them all at once?) the system runs out of memory, and there goes your OS.
Removing all language packs (other than your region default) fixed the issue on my system - though reapplying the language packs I needed was a bit time consuming.
How this bug escaped the lab is baffling - do none of the test systems have language pack installs?
Paris - because she can't figure out how this got out of test phase either...
I concur - my industry wide personal experience with one particular machine is that I installed 9.04 and after a year or so of updates and the occasional reboot, it no longer bothers me with update notifications! This leads me to believe that I now have the ultimate, final, fully maxed version of Ubuntu which never needs any more development so surely must be classed as The Greatest OS EVER.
Joking aside, I really should upgrade but I need to build up the enthusiasm to find all the drivers and support for the various bits of bobs inside the machine and then fiddle so that it works. Maybe once this darn project delivers I can set a weekend aside...
I've also had a mixed experience over the weekend. I likewise thought it was about time I applied the pack given that I'd heard nothing bad since the release. Nothing to do with WSUS, just manually updating machines at home.
Had to try the hack that was posted on the technet forums to get the machine starting again - fortunately it didn't need a complete re-build - it eventually backed-out the SP1 update.
I was greeted with this issue on Friday morning. It took the IT guy several hours (several computers, and several OSes) to fix it using the steps described at:
http://social.technet.microsoft.com/Forums/en/w7itproinstall/thread/1c9a7151-b48c-4a98-aae7-a4b82682ea8e#bcabda57-7338-499f-aee2-d708e76df315
Apparently, if you have the admin account turned on, you can get to the recovery console from the diagnostic GUI. I was not so lucky..!
Can anyone explain how you could do 90% of the stuff you do nowadays on an operating system that took up 5MB (a bit more if you include browser, media player and expansion card drivers)?
I tried Kubuntu -really nice-, but to change some basic options you need to type in 10 lines of text.
Another alternative is AROS (a free modern x86 Amiga OS) still needs work though.
If MS want to sell me another copy of Windows they should go back to square one and write a lean streamlined OS for 'home' users. What they do for other users doesn't concern me
"Can anyone explain how you could do 90% of the stuff you do nowadays on an operating system that took up 5MB (a bit more if you include browser, media player and expansion card drivers)?"
The answer to your slightly confusing sentence (I can only presume you typed something wrong) is that you CAN'T do 90% of the stuff you do nowadays on a 5MB (think MS-DOS) operating system. No GUI, no internet (just drivers alone for the tons of modems out there would be more than 5MB, not to mention NICs if you use a separate DSL/Cable modem instead of a simple dialup). Of course, a browser that is NOT lynx or the like would run at least 5MB too.
A fuzzy, feel-good interface (even OSX) takes up a lot of disk space. Functionality takes up space as well. The OS is likely "bloated" in your opinion because you don't even use half of the functionality that is available through menus, let alone any of the more hard-core functionality (when did you last use Group Policies? [just to name one]). A Linux install takes over 1GB (usually around 4GB for more common rollouts). I definitely wouldn't label Linux as a "for everyone" OS just yet. I feel 100% comfortable, however, having Win7 stuffed on machines though. No need to walk the parents through command lines or VI editing format-reliant config files.
Sort of. The final installations sometimes spilled over, though it should be remembered that Windows 3.1 usually took 6 floppies and possibly another four for the Workgroups version. I suppose it really depends on what you are going to do with the OS, whether you actually need a GUI and, if you do, what bits are of any use. You have to remember that Windows, Linux, Amiga OS, heck even RISC OS which is my own stamping ground, are built for general purpose use and will have a lot in them that you may or may not use.
The difference, I find, tends to be in the fripperies that go with the GUI to tart it up to appeal to those folk that care about such things and all the hand-holding wizards and such, some of which aren't much use. For example, how many users out there mod their Windows 7 installations to switch off the nags and the fancy screen bits to grab extra speed and reduce resource hunger? I still tend to use KDE3 on Linux because, up to now, KDE4 has been too hungry when doing video playback amongst other things.
Yet I still like to do my DTP work on RISC OS. And that's a tiny system compared with the modern stuff, GUI included. It's all horses for courses really.
... but I know that tinycore linux can get you a working desktop with firefox in under 10.
http://www.tinycorelinux.com/
Linux has the advantage of being monolithic, and is therefore pretty fast.
I don't think that resorting to something like Amiga would be the long term answer.
"I tried Kubuntu -really nice-, but to change some basic options you need to type in 10 lines of text."
Really? You must have screwed up the install pretty badly for that to be your last resort! I admin Solaris boxes day in day out from two Ubuntu desktops, the last time I typed in some command line stuff on my desktops was about 6 months ago to get some WAV files converted to MP3 so I could listen to some music on Saturday morning callout.
Can we drop this crap that Unix always needs command hacks, it's getting very tiresome and in 95% of the cases it's utter shite spouted by people who have never actually run a distro in their lives and simply want to put it down. If you don't like *nix, fine I have no problem with that, just stop ragging on it!
I think WIndows is a great O/S it has opened up tech for so many who would never have considered a computer before, and this from someone who uses Mac with OSX at home and Solaris/Linux at work, I hardly touch Windows these days and still think it's a fine O/S.
Part of the problem is that you may well be able to do everything in Unix or Linux from the GUI, but there are a lot of Unix/Linux 'gurus' who suggest that the only real way of doing something is through the command line. These are often the same people who call GUIs "point and drool" and slag Windows off, incorrectly, for having to do everything through the GUI.
Personally I use Win/Linux and OSx, I always have the command line open, but I am under no illusions that I'm not normal in this respect. There are a lot of others who still think that computing should somehow be difficult, if it's being done properly, a bit sad really.
that is indeed part of why people use command-line, (a similar phenomenon on windows is "registry hacks" instead of just checking the checkbox in options), but part of it is when you are wanting to walk someone through something, it's easier to tell them to copy&paste from the command line than it is to say "bring up this menu, then that menu..." ad infinitum. Windows server MSDN articles do the same thing; to enable a lot of functionality they'll tell you to type something into an elevated command prompt rather than click on the "install server role" option or similar
I run Win 7 hp 64-bit at home on two machines. A few days before the update was released I ran full system backup manually on both of them and then took backup off schedule. Happily when Redmond sent SP1 down the pipe to yours truly it installed in both cases without giving any noticeable problems. Had I experienced problems I could not find a work-around to I would have switched off Windows update, formatted the drives, restored from the drive image and then kept that monkey (AKA Windows updat) off-line until it was clear that a safe version was available. I've been burned before!
When they tell you to back up your system volumes (system and boot at least) before installing an update, they aren't joking. Having said that, my HP OEM tablet-laptop took it fine, too.
I think you can tell Windows Update to stop Windows 7 Service Pack 1 and still give you other Windows 7 updates, and that way you probably have everything safely updated anyway, but don't take my word for that.
My install of SP1 on my daughter Win7 pro running PC was a mixed bag.
On one hand, the Safari browser has slowed to a crawl and I had to install 5.0.1 to restore a bit of speed. On the other hand the "shut down" function which had ceased to work for some reasons weeks ago forcing me to use hibernate is now working again.
Looking around for the solution for my failure (1 of 2 machines at home worked) it does seem that it is x64 that is having all of the trouble.
Home edition went fine for me though but Professional failed (both x64).
After running through many of the suggested hoops ending in wiping my entire update history it eventually found an issue with some of the system files for which a system recovery would be needed. The machine has some deeper issues though, possibly due to a power outage mid update some time ago, so is due for a re-build anyway.
After various successes elsewhere I had just allowed SP1 through the WSUS. Worked on all but one of the machines and the failure on the 3rd isn't this one luckily enough.
64-bit Windows 7 Ultimate on development desktop with 64-bit Office (big Excel modelling) worked a treat, but 64-bit Enterprise version with 32-bit Office is giving me grief ... in permanent update cycle "Restart computer to finish update, installing, restart computer to finish update, etc." VM setups also giving me permanent loop of non working updates.
I let SP.1 install.
Machine was fine for a week.
Then broke badly, and there were no restore points.
Had to do two factory resets over the weekend.....the first left it unable to download files above 26meg without hanging.
As I was not aware that SP.1 may be the culprit, I let the machine update itself to SP.1 again.
It's like typing on a ticking time bo.......
I think it's pretty much luck of the draw.
To answer all of the "______ is better, never screwed up on update" (insert name of OS), I have witnessed and indeed screwed up myself OSX, Ubuntu, XP installations by following the prescribed update path. All OSs have their issues. The worst was an Apple that had to be sent back, update made the HDD disappear.
I suffered from a different error to the ones (above), namely a dreaded 'out of memory' error. Apparently (after some research) its well known to Microsoft- and the 'work-around' is to uninstall any additional language packs you may have installed. Not only is this troublesome- but the cryptic manner in which you have to find a solution, is annoying as hell......
I also had the hanging on update 124,556 of 300,000 (odd) updates on startup, and a myriad of other issues- that necessitated a fresh install......
Microsoft should be ashamed to have let this out the door- if we did similar with our software, we'd have our business users baying for our blood.......
I work in the UK business support dept for a PC hardware manufacturer that starts with D and rhymes with Hell.Since SP1 came out even home/small business users users have had the same errors described. I think its been rushed out of the door too fast without the necessay testing reqd.
Maybe this is a clue:
of our 7 machines, 2 run 32bit win 7 pro - they upgraded OK
of the remaining 5 running 64bit win 7 pro, 2 upgraded OK
the ones the failed all had extra language packs loaded - that is the ONLY difference in software packages.
The language packs loaded were Chinese Simplified, Chinese Traditional and Japanese, all to support overseas projects. Our standard installs have only Australian English.
I have read on a few other sites that it looks like the additional language packs might be the culprit.
We fixed our machines using a WinPE boot CD and removing the 0000000000000000.cdf-ms statements from %systemroot%/WinSXS/pending.xml. Apparently MS tells us this will cause extra problems, but I only found that bit after I had 'fixed' our machines.
Thanks MS. All the goodwill I had towards Windows 7 and SBS2008 evaporated in one quick update.
I hate MS again now. Hating MS is cool.
We're 100% 64-bit Win7 here. Tested the SP1 on a bunch of machines for last couple of weeks with no issues. Started rolling it out slowly across the company and started seeing the C00000034 fatal errors on a few machines. Thus far it's running about 15%-20% failure rate with *nothing* in common we can find yet. Machines created from the same default image running the same patch levels and same software will glitch and some won't. Total crapshoot with every install.
For those of you hit by this, the MS TechNet article on booting into startup repair, getting to a command prompt, and manually editing the %windir%\winsxs\pending.xml file *DOES WORK* every time. It gets the machine bootable and *appears* to back out the SP1 changes. However, if you do a check on My Computer > Properties you'll see it says SP1 is installed even though the SP1 installer says it failed. I haven't figured out whether the SP installed or not and I don't have time yet to go crawling through DLL versions to find out.
I've tried manually re-installing the SP1 from the downloaded (non-WSUS, non-Windows Update) offline installer on a few afflicted machines. Every one of them dies during the initial install claiming that some "windows component" needed for install isn't present. There's a "helpful" link in the error that says something like "why is this happening." It leads to an MS web page that says some OEM installs omit certain Windows Components (it doesn't say which ones) and -- surprise surprise! -- says you can *easily* fix the problem by RE-INSTALLING WINDOWS. Pardon me while I go strangle the idiot that came up with this web page.
If anyone can find some commonality in these failures, I'm all ears.
Whilist some problems are likely to sneak through in a major software change, the fact you have a 15%+ fail rate indicates Micro$haft is simply not testing is products in any meaningful way.
The suggestion here is never deploy MS SPs until you have to, and even then, do your own integration testing, and backup everything.
Praise be to the geek who worked this out (first reply in the thread):
http://answers.microsoft.com/en-us/windows/forum/windows_7-system/fatal-error-c0000034-applying-update-operation-282/bd99ead9-ece9-48af-9ad8-3c0377d01301
My laptop went from brick to fully working in less than 20 minutes, but woe betide anyone who's even slightly scared of getting their hands dirty with an OS. And it was the FIRST time in as long as I remember that I just OK'd a Windows update to install without looking at it first! >_< It was all going so well with Win7...
Well i didn't have a problem with mine!
running win7 home premium, but i wonder if the backup still corrupts DVRr/w's a few weeks after you forst use it?
i am loathed to bugger another one again, having tried two different laptops and numerous dvd's and all end up with same problem although if formatted they work fine for anything else !
I'm the guy that reported this to the register, and to all those that say why didnt people test before approving the update in WSUS need to understand that the majority of people affected by the c0000034 fatal error are people with SBS2008 servers and OEM factory image machines. In this situation, SBS2008 is preconfigured to deploy service packs automatically to clients. As soon as it became a critical update, thats when the problem went widespread. The support company i work for (Pyranet UK Ltd) have several clients with this setup, and its caused a fair few issues for them. MS need to sort this out!!!
Or did they just assume that Microsoft's choices were the right ones for their company?
Although a rank amateur with WSUS having installed it on our SBS setup I went into it and looked at the settings and set it so that no update was installed without someone saying Yes to it. And then after the list of updates got updated each month I'd go through and OK them one at a time. Anything related to a critical system, eg Exchange, I'd hold off on for a week to see if any bugs crept out of the woodwork.
The only change I've noticed on my machine is that it takes considerably longer to boot up. The extra time comes after the BIOS gubbins and before when it says "Starting Windows" or whatever. The screen goes black and I get an old DOS-style flashing "_" cursor at the top left of my screen that stays there for about 10 seconds before Windows starts loading. Anyone else had this issue?
http://social.technet.microsoft.com/Forums/en-US/w7itprogeneral/thread/608ecca8-b815-4ff6-8f3c-a828518434a7
INSTRUCTIONS:
01. Reboot your computer while it's starting up.
02. When your computer starts up again, choose the option "Launch Startup Repair"
--> PIC: http://notebooks.com/wp-content/uploads/2011/01/Recov-1.jpg
03. When the Startup repair starts, click cancel.
04. After you click cancel it will show a box. Click "Don't Send"
--> PIC: http://i52.tinypic.com/xgjriw.png
05. Click the link "View advanced options for recovery and support"
06. In the new window click Command Prompt at the bottom.
--> PIC: http://i51.tinypic.com/50imu8.png
07. In Command Prompt type this and press enter: %windir%\system32\notepad.exe
08. Notepad will open. In notepad go to File-->Open.
09. Change the type of files notepad views from .txt to All Files (see pic)
--> PIC: http://i51.tinypic.com/35nd74z.png
10. Now in Notepad, go to C:\Windows\winsxs\ (or whichever drive Windows is installed on)
11. In that folder, find pending.xml and make a copy of it
12. Now open the original pending.xml (it will load really slow because the file is huge)
13. Press CNTRL+F and search for the following exactly: 0000000000000000.cdf-ms
14. Delete the following text (yours will be a little different):
--> PIC: http://i54.tinypic.com/adzpzp.png
Your PC might not have all 3 sections of code (, , ). Just make sure you delete section "Checkpoint" and whatever other sections have "000000000000000.cdf-ms". They will be right next to eachother.
15. Save the file, close notepad, close command prompt, restart your computer.
Once your computer starts up, do a normal startup (it may stall for 5-10 minutes at the "starting windows" screen, but leave it going) and the Service Pack will install some more stuff and restart a few times and then everything should be working! For some people, it reverts everything and cancels the service pack installation. For other people, the service pack installation completes. Either result is fine.
I posted this link earlier, but it's "buried" back on page 1. This is the procedure which got my machine working again, +1!
Side note: the Command Prompt (step 6) is only available to the Administrator account. The "Administrator" account was disabled on my machine as I the sole user and had admin privileges. Ultimately, the file had to be edited from a different machine.
Hope this helps someone...
I think El Reg should prevent comments on any articles which points out flaws in any personal computer OS, these comment threads always end up at the same places...
<start>Microsoft bad</start>
<middle>Linux great</middle>
<end>It's just personal preference</end>
Obviously there is a lot of fluff and insults in between but they always follow the same structure.
"As an ubuntu user since version 5.10.... #
...and various BSDs before that, I'd like to point out that all operating systems have bugs, even Linux."
NEVER!!! Linux is perfect!!11one.
But seriously, I am a little disappointed in Microsoft releasing these machine-killing service packs -- they did it before with an XP service pack, where certain OEMs installed an Intel driver on AMD systems, it worked fine until I think SP2 (or maybe SP3?) was installed then they'd drop dead. Now, again with the 7 service pack.
------------------------------------------------------------
@Big Bear, re: "Joking aside, I really should upgrade but I need to build up the enthusiasm to find all the drivers and support for the various bits of bobs inside the machine and then fiddle so that it works. Maybe once this darn project delivers I can set a weekend aside..."
You probably don't need to. I've run updates from 8.04 up to 10.04 then 10.10, and I didn't have to reinstall a thing. My parents even have the Samsung-provided driver for their CLP-510 (it had SLIGHTLY better color matching than the open source driver) and THAT kept working after going from 8.04 all the way to 10.10. If you don't have an option showing to upgrade, you run "update-manager -d" and then you do. Of course I'd still wait for your project to deliver, doing an upgrade in the middle of a project could be a bad idea.. but I think you'll be surprised at how smooth it goes.
---------------
RE: Everyone who brings up the Linux command line. I have one response: Regedit. You seriously cannot pretend that starting regedit, going in the registry to "Local Users"->"Blah de blah"->"The fourth choice", inserting a string object called "foobaz", and putting in "{ABC12-2345-CDE-DEADBEEF}" for foobaz is in any sense easier than having to go open /etc/foobar with a text editor and put "somejunk=1" in there. The reality is, in this day and age, normal Linux users don't have to crack open a text editor, and most Windows users probably do not crack open regedit.
Turn of WSUS??? For real? That comment is a little extreme. If you are using WSUS just tell it not to install the Service Pack.
When I heard W7 SP1 was coming I immediatly turned off deployment of service packs for Win 7 until I know it is not going to crash half my users PC's... thankfully I did since it appears to have a negetive impact on OEM installs....
I am astounded that you would simply tell people to turn off WSUS completely.... sheesh
If it ain't broke don't fix it DOES NOT APPLY TO SOFTWARE, there are bugs in code that allow exploits. You should keep your software up-to-date for that matter.
Of the many OS's I have used in time, I can say the following:
Windows NT4: seen BSOD after SP4 and SP6 installations.
Windows 2000: seen BSOD with SP3 and SP4, iirc.
Windows XP, to this day, I have witnessed BSOD's after installing every SP (SP1, SP2 and SP3), not on the same machine but on multiple machines, with various configs, ... I used to work in technical support.
Linux (user since 2001: Suse 7.0-9.0, then Ubuntu dapper drake, now Debian) update issues mostly with Ubuntu, I switched to Ubuntu when automatic updates in Suse changed my glibc without updating my shells := screwed. With Ubuntu, I have had "many" issues upgrading, the worst being the idiotic switch to ash from bash so I switched to Debian, happy ever since.
Mac OS (user since 2001: two kernel panics with 10.2 caused by Norton, experienced in the lab - I was working for Norton, none since)
Solaris: One thing about Solaris is, you would spend more time finding the update on sun.com than anything else ... never had issues after installing the updates, once I found em that is ... ;-).
FreeBSD: never experienced any issues updating.
Needs about 6 GB of free disk space to install, yet I only had 1GB free. The feedback from Windows Update was atrocious (just an opaque error code with no useful help link), so I had to manually install the SP to find out what was wrong. Now I have big hassles repartitioning my C: drive :(
If you have automatic updates set up at all your workstations there is no way to avoid installtion of SP1 after it has downloaded. I had 6 PCS affected with that fatal C00000034 error on the 9th and at the time Microsoft was NOT even aware of the issue. Thank GOD I found a fix that day here:
http://social.technet.microsoft.com/Forums/en/w7itproinstall/thread/1c9a7151-b48c-4a98-aae7-a4b82682ea8e#bcabda57-7338-499f-aee2-d708e76df315
I was able to bring stations back up without losing anything.What I didnt know was that the restore points I had set up were erased because of this update. Ironically the remaining windows 7 stations installed with no issues but appear to have slowed down performance the next day.
Businesses CANNOT AFFORD to be down and Microsoft needs to make sure they test thier updates thoroughly. We would need a crystal ball to tell us when we shoud turn on or disable WSUS.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
