Why you don’t want to edit your pending.xml to resolve 0xC0000034 issues


I know that one of the workarounds everyone is using is editing the pending.xml file and rebooting the machine.  This is seemingly working for everyone that the SetupExecute registry value does not.  I wanted to give a brief rundown as to why you only want to do this as a last resort.

During service pack installation, we populate the pending.xml file with all of the files and registry values needed to install a particular update.  Service packs are special in that they are broken into critical and non-critical transactions to allow us to recover more quickly and reduce the no-boot window that could occur during installation.

During system shutdown, we process all of the critical transactions first and then the non-critical transactions.  If we fail processing the critical transactions, the service pack will just fail and rollback.  If the critical operations succeed but the non-critical operations fail, we attempt to process them on reboot using Session Manager (smss) and the SetupExecute registry value.  When the system reboots and reads the SetupExecute key, it retries installation first and if that fails it will roll back the Service Pack installation.  Deleting the registry value tells smss to not try and run the poqexec.  It should be reattempted again during startup processing or fail outright.  So effectively deleting the registry value breaks you out of the install fail reboot loop that the machine ends up being in.

Additionally, the pending.xml file has a checkpoint value that tells Windows where the critical transactions end and the non-critical transactions begin.  When you delete the checkpoint value in the pending.xml, its effectively marking everything in the pending operation queue as critical.  Because your machine has already rebooted, Windows thinks it has nothing to do and just boots normally.  The problem with this is that because there are still operations that need to be processed that will not get processed and this could potentially leave the machine in an even worse state. Doing this should be an absolute last resort. The best thing to do here is let the failure occur later on so a rollback can take place.

–Joseph

Comments (222)

  1. Anonymous says:

    In most of the instances I have worked on with this, using the registry to rollback and then reinstalling the standalone package allows the service pack to install properly

  2. Anonymous says:

    If you managed to remove and reinstall the service pack than I would say you're probably fine.  You can delete the pending.xml  files from the machine (although what I would do is copy them off of the machine and then delete them, just in case).

    Thanks for the information.

  3. Anonymous says:

    Phil;

    The KB link is in my main C34 article.

    So you used method 2 and recreated the registry key and the machines did not roll back?  Can you upload your CBS directory to a skydrive or share I can hit?

    –Joseph

  4. Anonymous says:

    Regback is basically the same as removing and readding the key.  The difference would be whether you saw the service pack rollback or not.  If it rolled back, then you're probably fine.  If not, you might still have payload in a pending state.

  5. Anonymous says:

    NP Mike, glad to help out as much as I can.

  6. Anonymous says:

    @rfx, Glad that worked.  I would run the CheckSUR utility against your machine to make sure its in a good state.  It's KB947821.  Let me know if thats clean or reports errors.

  7. Anonymous says:

    Christian;

    I totally understand and you have to do what you have to do.  I'm working on a method to get machines that have used the pending.xml hack back into a good state.  Now timeline on when that will happen.

  8. Anonymous says:

    Drew;

    I think it would be irresponsible on my part to say what might be causing it without any real data.  I know of ways to make it happen, but thats manufactured data and not real.

  9. Anonymous says:

    It takes time to triage these things and find out what everyone is doing.  I cant back test every method someone uses to recover from a failure and posted this as soon as I had more information about the problem.

  10. Anonymous says:

    Got'em, thanks!

  11. Anonymous says:

    Gotcha.  Let me see if I can figure something out.  I'll let you know here if I do.

  12. Anonymous says:

    @Mike;  I cant say in regards to the repair piece, I know its being discussed but I dont know what the outcome of that is going to end up being.  Rebuilding should fix the issue or at the least give you a better way out of it with the script if you were to hit the problem again.  Basically, if you dont need to do anything now, I would wait.  If you find that you need machines in a hurry and you're having other updating issues, then I would recommend that you rebuild those machines with either SP1 integrated media or however it is you would normally rebuild a machine post service pack release.

    Let me know if that makes sense….seems like a lot of words 🙂

    –Joseph

  13. Anonymous says:

    @ Andy;

    I'm not doubting that.  But what I am saying is that when this issue hit, we went back and set this up many times and couldnt get it to happen until we started manipulating Windows to produce the error.  For a long time I was actually looking for a repro via the newsgroups so we could debug a machine in state because we couldnt get it to happen at will.

  14. Anonymous says:

    Thanks for the info Flo.  I'll look over the ones I have seen and see if that has any commonality with what I have.  I appreciate the information.

  15. Anonymous says:

    @Dick;  Shutdown, while common in some of these cases, isnt the only thing that causes this.  I have personally tested the shutdown+WSUS+SP1 many times over and not been able to reproduce this in house.  As I've stated above and elsewhere, if you are at a last resort, then do the pending.xml hack.  Just be aware of what you're actually doing to the machine when you do that and the state that its putting machines into.

  16. Anonymous says:

    @Joe, Thanks for the info.  I dont need logs right now but if I do I will let you know.

    @Dick;  Some machines that have the pending.xml applied to them are able to remove the service pack and reinstall using the standalone.  Most are not however.

  17. Anonymous says:

    @Bret:  You can use the hack if you're in dire circumstances, just be aware of what its doing.  People are going to have different behaviors depending on where the servicing operations were when the problems occurred.

    @Phil:  Aside from a small percentage of machines, the registry value always forces a rollback.  I have yet to see a post registry change result in the appearance that SP1 is installed but have seen some that people have gotten config errors post reboot.

  18. Anonymous says:

    @chucko

    It's definitely related to the torn state although I'm also aware of the other issue caused by Driver Sweeper. The fact that the driver is missing in this instance is because SP1 updates it and replaces the folder with a different version. When trying to reinstall SP1 on a torn machine, the driver has already been replaced once and so the install fails.

    (apologies if this appears twice)

  19. Anonymous says:

    @Mike, thanks I appreciate the patience on everyones part on this

    @Ellak, yes, it's something myself and others have looked into but I have seen English only systems in this same state so its hard to say if its a red herring or an actual contributor to the problem.

  20. Anonymous says:

    @Mike

    Have just completed an Upgrade install on a machine that used the pending.xml fix and can confirm that the machine is back up and stable. As Badazzwinger said, you have to then reactivate certain windows features such as IE8, Windows Search, .Net3.5 etc. No idea why that happens and still don't think this is the ideal fix but at least it works.

  21. Anonymous says:

    Okay, so I've got a fix that works for me. If anyone would like to give this a go on a torn machine and post back their results that'd be really helpful in working out whether this fix is specific to me or more common.

    Here's the steps I've taken today in the order I've taken them. Technically the first two attempted installs of SP1 aren't necessary but as they pointed me in the direction of getting it working I've included them.

    1. Attempted to install SP1 from DVD on a torn PC. – Failed.

    2. Noted that the CBS.log reported the folder %windir%System32DriverStoreFileRepositoryatiilhag.inf_amd64_neutral_951c1812f542740a as missing

    3. Extracted that folder from the install.wim on the original installation media using 7zip

    4. Booted to recovery console and copied that folder back using xcopy

    5. Rebooted into windows and set permissions and ownership of folder and files to match that of similar folder atiilhag.inf_amd64_neutral_0a660e899f5038a2 (barring inheritance as I couldn't be bothered to do each file. I'm sure it could probably be done more easily using CACLS or something).

    6. Attempted to install SP1 from disc. – Failed with different error.

    7. Noted that service pack (Package_for_KB976932~31bf3856ad264e35~amd64~6.1.1.17514) reported as partially installed in CBS.log.

    8. Ran dism from elevated command prompt with command dism /online /remove-package /packagename:Package_for_KB976932~31bf3856ad264e35~amd64~6.1.1.17514

    9. Rebooted when prompted

    10. Installed SP1 from DVD media. – Success!

  22. Anonymous says:

    I think both of you are right.

    Susan is right to expect an admin to own their environment and ensure only the changes that happen to that environment have been fully tested prior to doing so.

    Phil is right in expecting the service pack to perform as it has during beta/release cycles.

    I look forward to the logs, and please, keep an open discussion going.  This is how things get fixed in the end (and logs <G>)

  23. Anonymous says:

    @Mike;  Nothing yet so if you need the machines, start the rebuild process.

  24. Anonymous says:

    @Phil;  if you can get me logs from the machine then we can look over them

    @Joachim;  Just trying to be helpful here, if you find it useless, dont read it.  I think its pretty useful to know what is actually happening on your machine when you follow a specific course of action such as editing the pending.xml file.  If I dont post this, everyone would think they were in a "good" state when they clearly arent, if I dont post this then everyone would think we're trying to hide something.

  25. Anonymous says:

    @Mike;  we're still actively working on it

  26. Anonymous says:

    @Rahul:  It's a pretty big assumption that these are all the same issue.  Our error handling code for CBS is flat.  Meaning that this particular return value happens for a reason but the reason it happens can vary.  It's not any different than other error handling mechanisms in code. For example, I can make this error happen at will by doing a few specific things to an image, that doesnt mean that its the same as what is happening to you but it will return the same error to me.

    I was one of the beta engineers in enterprise for SP1 and I worked on exactly one issue related to this error code, just to give you a frame of reference.

  27. Anonymous says:

    Just make sure you install the service pack by itself and restart the machine and you should be fine.

  28. Anonymous says:

    @reader;  if you're in a torn state, rebuild.  I dont have a time table on torn state

    @rfx; glad things got sorted out for you.

  29. Anonymous says:

    @Andy;

    Throughout many, many tests, this issue didnt occur internally over many testing cycles.  In fact, to get this to reproduce internally we had to physically manipulate machine states to get them to C34.

  30. Anonymous says:

    @Mike

    No problem. Just got the post up now at kelvinaston.blogspot.com/…/possible-fix-for-torn-sp1-machines.html – hope it's helpful.

    Also apologies to everyone as I've missed a tilde out of the dism command I've posted in these comments. The correct command should be:

    dism /online /remove-package /packagename:Package_for_KB976932~31bf3856ad264e35~amd64~~6.1.1.17514

    Thanks to @dickfrey for pointing that one out!

  31. Anonymous says:

    Joseph just an FYI follow-up here that may or may not have been a factor.  All the machines were 64-bit and encrypted with Bitlocker.

  32. Anonymous says:

    Dick;  Sorry, but thats just not the case.  It might be what you're seeing but its not something that is 100% reproducable.  Trust me, I've tried.  As for a fix, as I have said before its something we're looking into very actively.  I hope to have more information soon.

  33. Anonymous says:

    It's not just the 64 bit systems that have this issue.  One of my company's clients has about 80 computers and all of them are 32 bit.  Out of the 80, 20 of them took the SP1 installation without any problems and about 5 so far have gotten the "Fatal error c0000034 applying updates operation XXX of XXXXX (0_000)" and I have been applying the pending.xml fix.  This morning I had the client contact me because the error came up and we performed the pending.xml fix but this time we got the error again but with a slight change.  The (0_000) had changed to (prig…) this time so I'm not exactly sure what had caused this.  I tried the registry fix but kept getting "ERROR: The system was unable to find the specified registry key or value."  The first time if happened, I was doing the "Reg load HKLMBaseSystem C:WindowsSystem32configSYSTEM" so I changed the C: to D: and it worked but then the next line "Reg Delete HKLMBaseSystemCurrentControlSetControlSession Manager" /v SetupExecute" gave me the error statement again.  So where do I go from here??

  34. Anonymous says:

    @Ray;  Yes our solution is tested and verified to work on in state machines.  Our script will not result in a torn state and will allow the service pack to install properly.

  35. Anonymous says:

    Terry;

    The entire point of this entry was to not edit the pending.xml, if you've done so, you've put your clients into a torn state and you're not going to be servicable.  We have workarounds and guidance for this issue through WSUS server pushes available both here and on the external KB now for working around these issues when and if they occur.  For the systems in a torn state you can either rebuild them now and make them servicable.  We're looking into other options for torn state but I am not promising anything.

  36. Anonymous says:

    Yes we published that last night

  37. Anonymous says:

    @Rahul;  Still working on it, will have something as soon as I can.

  38. Anonymous says:

    @Rick;  No, it doesnt imply that we can reproduce the C34 at will.  My guess would be I could make it happen if I deleted the checkpoint file prior to a reboot, but thats a manufactured example and not a real one.  I honestly dont think that WSUS has anything to do with this personally.  I think its image related in some way.  I've seen this same error occur on non-WSUS installs.  Unless there is something in the way WSUS rolls out the patch what would be different than WU/MSU packages, that would be the only difference.

  39. Anonymous says:

    @joscon

    I believe that I have a fix to get out of the torn state. I've only tested it on one computer as yet and I'm hesitant to post it here as yet in case it causes other problems for people. Is it okay to email you my procedure to see if you think it's harmful?

  40. Anonymous says:

    @JMPNY:  Do you know if SP1 shows in add/remove to where you can run the uninstaller?  If not, I might have to see of a way to manually try to remove it.

  41. Anonymous says:

    Having just rebuilt a machine from scratch and monitored this folder, it seems that SP1 updates the driver from the 951c1812f542740a version to the 0a660e899f5038a2 version as part of the install. Attempting a reinstall without the original version present obviously fails. Those machines on which I've applied my fix and then SP1 now have both versions of the folder. I'm not sure what, if any, consequences this might have in the future.

  42. Anonymous says:

    We did all the testing for this before and this did not show in our testing, which is why we deployed.  No that is the problem it does not show in Add/Remove programs (my note indicated that) but th emachine thinks it is installed somehow.

  43. Anonymous says:

    @chucko;

    "IMHO, you seem to be missing the point that others are trying to make."

    –joscon: I'd say I am actually one of the few who actually understands your position.

    "It is poor practice IMHO to look at call volume as the primary indicator of how serious a problem is."

    –joscon:  I 100% disagree with you on this.  When you have an install base of 400-500mil clients, call volume is a great indicator for how serious a problem is.  I dont call my doctor unless something is wrong, people do the same for their enterprise installations.

    "You seem to be wanting to dismiss the problem because in your mind there are few PC's affected."

    –joscon: Dismissing the problem would not entail me responding to your comment.  I have actually done quite  a bit here to give you a voice internally and make sure others know the pain you've seen based on this issue.

    "We all know that the numbers of PC's in a "torn state" is far higher than you think it is."

    –joscon:  This is a pet peeve of mine.  You dont know what I think the number is so please dont hazard to guess.  Again this goes back to scale of deployment though.  My personal opinion of this though is that its actually the opposite, I think many people think this has hit a much larger base than it really has.  I have numbers from a lot of people worldwide and they are not out of line with other service pack installs.

    I'll say one other thing here though.  I do all of this on my own spare time to attempt to give others a little more understanding of how the servicing stuff works and a platform to voice their complaints.  A lot of those have now started to turn against me though, and I wont really have it.  If you want to complain, cool.  If you want to jump to assumptions as to what I am and am not doing in the way of helping you, you're free to but I will most likely just kill off the blog and do something more productive with my time.

    –Joseph

  44. Anonymous says:

    @dickfrey

    That’d be awesome, although I’ll settle for a job/consultancy fee instead 🙂

    It’s now worked on half a dozen of my machines and they’re all once again receiving updates. So, can anyone out there with a torn machine please at least try the first couple of steps and see if they’re missing the same file please?

    1. On a torn PC try to install SP1 from the full file download
    http://www.microsoft.com/…/details.aspx

    2. When it fails, check %windir%LogsCBSCBS.log. What’s the error? Is it complaining about the lack of the atiilhag.inf_amd64_neutral_951c1812f542740a folder?

    Thanks to anyone who can help

    K

  45. Anonymous says:

    @Drew: Critical in this case refers to the transaction from a servicing perspective, its more of an internal usage term in this context

    @Rahul:  I didnt disagree with you then and I dont now about the benefits of repair.  I also agree that any problem is a big problem when it comes to Windows because of the install base (and the passion of our userbase <G>).  We're working hard to see what's going on here.  I have my own suspicions but I'm keeping them to myself for right now.

  46. Anonymous says:

    @Leen;  I'm assuming you're speaking about CheckSUR.  CheckSUR has nothing to do with the SP1 release.  It was just re-released to work with SP1 installations.  If you have errors running CheckSUR then you have issues with the servicing mechanics on your machine and they arent related to SP1.

    SP1 does not run the CheckSUR utility before the installation because its an OOB tool that is used to detect corruption in the servicing directories on your installation.

  47. Anonymous says:

    @Spextor; Yes, SP1 is safe to install.  Please see the details in here for steps to avoid the problem during installation: blogs.technet.com/…/new-information-on-error-code-0xc0000034.aspx

  48. Anonymous says:

    On these machines that are losing the restore points, are there any drive space problems on the machines?

    @Nick: I understand what you're getting at but honestly, its not practical.  You could always parallel install the machines and move data using USMT to get them back up.  You could even do this with integrated SP1 media to save the SP1 headache piece.

  49. Anonymous says:

    Drew;  I havent seen an occurance yet where a reinstall hasnt allowed the service pack to work.  I'm sure it exists though.

  50. Anonymous says:

    Just fyi, we have 22 systems in this "torn state" (out of around 125 that WSUS deployed updates to the evening of 3/17/11).  The next morning, we scrambled to get the 22 user's up and running asap — they, and their boses were not happy.

    We would also like to see a supported fix to this situation vs. re-imaging the affected systems.

  51. Anonymous says:

    @Kelvin;  Sure, send it my way and I'll have a look.

  52. Anonymous says:

    Joseph we tested and retested SP1 from Beta to release and never had an issue before with WSUS.   So when we decided to pull the trigger and let the update proceed we were amazed to have 50 machines inoperable when folks came in to work in the morning.   That being said the only fast way to get everyone up and running was to use the pending.xml hack.   Fortunately we were able to stop the rest of the machines from deploying via WSUS or we would have been in critical situation.   The fact that we have the machines running now is ok but the machine is confused.  If you do an about it says SP1 is installed but it is not in Add/Remove programs.  An attempt to rerun SP1 from the file gives an error that it cannot proceed because compnents are missing.  So our issue is we have a bum install that we would like to cleanup.  Any thoughts?

  53. Anonymous says:

    @ripdevil: we wouldnt support any image that was created using utilities like that.  Microsoft does NOT support the integration of a service pack into an image.  I'll leave the comment as more of a warning to everyone than anything.  We've already seen other such utilities cause issues with SP1 installs

    @gmfile: You can attempt the installation again, yes.  Let me know the results

    @Brian:  I've actually been trying to reproduce the restore point failure because while I know it exists in the wild, I cant make it happen here.

  54. Anonymous says:

    @Mike;  Yes, we're still looking over other options for the pending.xml portion of this.  I dont have a timeline on when something might happen.  If you need the machine servicable now I would recommend that you rebuild/parallel install those.

  55. Anonymous says:

    @MIke, as WinDose says, the method does not work for torn states.  I'm still looking into ideas for that one.

  56. Anonymous says:

    @dick; Nope, thats just a temp directory

  57. Anonymous says:

    @Mike

    I've done it on one machine so far. I can't be 100% sure that I used the pending.xml fix rather than the SetupExecute fix as I've been treating them both as being the same and my brain is adled from the week's work on this and working on so many machines.

    I'll run it again on another machine that I can be sure was a pending.xml and get back to you with confirmation.

  58. Anonymous says:

    @rfx;  If thats the case and you dont have any restore points available then your best bet is to use the revertpendingactions flag in DISM.

  59. Anonymous says:

    @Mike;  The main detrement is the inability to patch.  So, in the torn state, you wont be able to install any updates properly including critical ones.  

  60. Anonymous says:

    Susan, thanks for the tip. I, of course, thought everything waited for approval. Now I know. Thanks again

    jtm

  61. Anonymous says:

    @dickfrey;

    I get that you're angry, and appreciate your feedback.  But, and I know people are going to hate this, let's look at it another way.

    –SP1 releases

    –We see a few issues internally but honestly nothing crazier than other SP releases (I tried to document the ones I was seeing)

    –C34 issue hits when the WSUS version of SP1 releases

    –We have a small increase in calls at Microsoft on the first day that this hits but those quickly trail off

    –I post what we know here and what I know on the forums (which is where everyone went)

    –thiswoot's workaround is posted

    –The day after that was posted I posted about why you shouldnt be doing it (aka, this post)

    –Microsoft publishes a workaround KB and script within a week of the issue being seen

    All in all, I actually think that this is pretty good support for an issue that was mainly seen on the newsgroups and here.  Had everyone who had seen this C34 error called in the day they got the error, it would most likely have generated an even faster response for the script workaround to alleviate these issues.  I've said it before but I cant fix what I cant see.  When you dont call me for help, I dont "see" the problems you're having.  And when most of you dont call in, then the company as a whole cant see that.

    I know a few can speak here that I personally made myself available almost around the clock for a couple of weeks to try and get as many answers for the community as I could, including hopping on the phone with some of you.  

  62. Anonymous says:

    @Rahul;  If you have a machine in a torn state, then no, there arent other options available right now if you need the machine in a servicable state

  63. Anonymous says:

    @Mike, sorry, thought I had answered that already.  No, you cant reverse the pending.xml workaround once its been done.

  64. Anonymous says:

    @Kristian;

    Because I dont have a machine thats broken in the same way that the groups appears to be, I cant validate this as a solution.  If more of you tell me that it is working for you, then we can look into it as a possible solution.

  65. Anonymous says:

    So just as an update, I personally am pushing to have the torn state fixed and am using the comments/frustration/etc posted here and sent to me personally as one of the means to have this done.

    I understand all of your points regarding how we gather data, I'm simply pointing out the nature of the beast (in this case how we use call data).  Could there be a better way?  Yes, there is always a better way to do things.  But, for now I have to work in the system that I have.

    Currently the advice is to rebuild systems that have had the pending.xml workaround used on them.  I dont like it, I know you dont like it, but thats where we are at this moment in time.  I am doing what I can on my end to be your voice internally here and will let you know what the outcome is once its decided.

    –Joseph

  66. Anonymous says:

    You're welcome.  I appreciate your patience.

  67. Anonymous says:

    @Mike;  I'll speak from my perspective about Kelvins idea here.  If you have that entry in your CBS.log then copying the files from good media and then DISM'ing out the service pack should be all thats required as a test.

  68. Anonymous says:

    @hagman;  No you cant rely on the bitmap, if you used the pending.xml hack you're most likely in a torn state.

  69. Anonymous says:

    @hagman;  I dont have anything for the pending.xml torn-state's at this time.  The admin password you're mentioning wouldnt indicate a torn state to me but some sort of other issue.

  70. Anonymous says:

    @Mike;

    Nothing yet,sorry.  We are putting a lot of effort into this.  I've tried to ensure that all of your voices are heard when it comes to the effect this has caused the community at large and you can rest assured that everyone involved understands that.  I'm not exactly known around here for being shy, so I am pretty sure my points are getting across. <G>

  71. Anonymous says:

    @Mike;  They're still looking at options for the torn state issue but I dont have anything new for it right now.  Once I know whats going to happen one way or another, I'll let you all know.

  72. Anonymous says:

    I have done the same as Kelvin on 5 different machines which were "fixed" using the pending.xml fix. I also had to re-enter a Product code. One other item I noted was that under Turn Windows features on or off, None of the entries showed their check marks. The group item showed that items were set by having the square colored. The first machine I upgraded was then missing .Net 3.5, IE 8 and the search box in the Start menu as examples. Once I discovered this, I went into the Windows features and made sure all check marks were set as I wanted them then did the upgrade. The biggest thing doing the upgrade is the amount of time it takes. On my systems, 2 and a half to three and one half hours per machine. I still have seven machines to do. I have 27 virtual servers and 3 physical servers all running Windows 2008 R2. I did not allow these machines to use WSUS for the SP1 update. I used the stand alone installer for SP1 on all servers and had 100% success with this method.

  73. Anonymous says:

    @Mike

    By saying that the first two attempts to install are unnecessary I meant that the actual install attempts only pointed me in the direction of the problem rather than making changes to the system. The first attempt confirms that the atiilhag.inf_amd64_neutral_951c1812f542740a folder is missing and the second that SP1 is partially installed and should be removed.

    I'm in the process of writing a blog post of the full series of steps which I'll hopefully get up within the next hour (would have been done by now but I've been in meetings all afternoon). However, if you're sure it's the same problem on your machines, an edited version of the steps would be:

    1. Extract atiilhag.inf_amd64_neutral_951c1812f542740a folder from the install.wim on the original installation media using 7zip

    2. Boot to recovery console and copy that folder back using xcopy

    3. Reboot into windows and set permissions and ownership of folder and files to match that of similar folder atiilhag.inf_amd64_neutral_0a660e899f5038a2.

    4. Run dism from elevated command prompt with command dism /online /remove-package /packagename:Package_for_KB976932~31bf3856ad264e35~amd64~6.1.1.17514

    5. Reboot when prompted

    6. Installed SP1 from DVD media.

  74. Anonymous says:

    Same here, had the issue at 10 PCs, all DELL (Optiplex, Latitude) last thursday.

    All PCs, running the update automatically (there were at least 5 updates to install via WSUS), hung at next reboot with error 34. The Pending.xml hack worked fine for me (and for the customer).

    Another 11 PCs I updated in the following order:

    1. Install the 4 other updates (exclude SP1)

    2. Restart the computer (at some machines i even did not restart, just to test, if there is any change)

    3. Install SP1 via WSUS (doesnt matter, if it installes on shutdown, or directly from the Windows Update Window)

    4. All finished and rebooted well, except 1, where I need to install SP1 3 times. (one successful, but it was still offered; one failed; and of course, still offered, and the last time it installes correctly).

    So i assume, that the issue occurs, when the SP1 is installed with other updates together.

    Can anyone confirm this?

    Like other contributors I need a routine, to get all PCs fine again, because I do expect errors with any upcoming hotfix due to pending.xml "reconfiguration".

    Martin

  75. Anonymous says:

    Hi. Have been watching this discussion all week so thought it was about time I threw my keys into the bowl as it were…

    I've got about 30 machines in my office and about 80% have now experienced the c0000034 error. Some I got back up and running using the pending.xml edit and some the SetupExecute delete once I came across that. All of them though are in the 'torn' state and nothing I've tried this week has helped get them out of it (I've even started a blog at kelvinaston.blogspot.com to keep track of what I've done).

    Obviously I now have a note of the official fix for use whilst experiencing the error, but I'm not sure that's going to be much use now. Ah well, these things happen.

    The only thing that has worked for me is what I've mentioned in my last post – an upgrade install over the top using media with SP1 integrated. On the bright side, my test machine now has a complete install of SP1 and all software and settings have been maintained. I suspect it wouldn't have taken me any longer to reimage through WDS/MDT though.

    Oh, and as Mike just said, you need to enter another Product Key for this method which isn't exactly great.

    I'm not planning to roll this fix out to every machine just yet for a couple of reasons; as far as my users are concerned they're okay just now and, more importantly, I'm leaving for a long weekend in 2 and a half hours and don't really want to open that can of worms just this second. With a bit of luck there may be a better fix by the time I get back.

    Thanks @Joseph and everyone working on a fix for this.

    Oh, and just to be clear, there's no finger pointing from me – I tested it with a manual install on 10% of my machines and it worked fine so I just rolled it out. Who could have anticipated that there'd be a specific issue with installing it using WSUS/install and shutdown? Not me. So how can I blame anyone else?

  76. Anonymous says:

    @dickfrey;  I'll respectfully disagree with your suggestion regarding thiswoot.  While I know he/she was trying to help, this blog entry was written specifically to address the problems that doing those steps might have caused.

    I'd like to know more about who else is seeing Kelvin's workaround work though.  And we are hiring 🙂

  77. Anonymous says:

    Has anyone tried this yet? I have time, I can wait!!

    jtm

  78. Anonymous says:

    @Kristian

    Thanks for the update. It's all still looking good at my end so fingers crossed.

  79. Anonymous says:

    @PGL:  So you did the registry workaround and it didnt rollback?  See if you can uninstall the bits you do have in add/remove programs –>installed updates if you havent already

    @Susan/Nick:  I understand 100% what you're hitting, my point was that a repackaged version specific to this issue isnt a realistic option.  I'm not saying that wont happen because that choice would be way above my paygrade anyways, but I do know that the packaging of a Windows release takes a long time, so a special repurposed version of Windows would take at least that long, if not longer, to go through testing before we would even release it.

  80. Anonymous says:

    @Rahul;  None that I have heard of. My team is the one that supports HV CSVs so I would have heard something by now if it were a big deal.

  81. Anonymous says:

    I have four Lenovos w/ATI (SBS 2008 WSUS) that three made it and one is toast. Three certainly updated by "Shutdown and Install Updates". All four may have, but I believe the fourth was installed by "Install updates at 3:00 am. The toasted on was definately  "Shutdown and Install Updates". But then I have two that made it through.

    jtm

  82. Too little, too late. says:

    Now you tell us?!

  83. bret says:

    Okay, but if we don't edit the XML file, the system never boots normally, and is useless.  It never fails the install, or rolls back.  How do we get around this without editing the XML file?  I have several PCs at clients who have experienced this issue.

  84. Phil D. says:

    Well for one this bug should have never made it past Microsoft.  Two they need to provide a fix for us that HAD to do this because it was critical to get people back up and running.  Finally, even the official Microsoft fix of deleting the registry entry still screws the system up because it thinks SP1 is installed, but it isn't.

  85. Phil D. says:

    Forgot to give the website for Microsoft's fix.

    support.microsoft.com/…/en-us

    You can't use method 1 because the Service Pack wipes out all the system restores (nice feature).

    Method 2 says it rolls it back after you reboot, but if you right click computer and go to properties SP1 shows it is loaded.  If you look under add/remove it is nowhere.  If you try to reload SP1 manually you will get an error and you can get the code from the event viewer. "Service Pack installation failed with error code 0x800f0a13"

    I have about 15 machines so far that are like this.

  86. Phil D. says:

    It says it rolls back and then you get the message that the SP failed to install, but WIndows is confused after that and like I mentioned, manual install of SP1 will not work after that.

  87. Joachim Otahal says:

    This blog entry is something for the trash without undelete.

    It is useless to say "don't do this" without offering an working alternative if you have several bosses entering which say that they cannot work any more.

    It is useless to say "let the error occur and let it roll back" if that does not happen.

    It is useless to say "go back to last system restore point" if it does not exist since SP1 miraculously kills the restore points if that error occurs (or it kills them before).

    This blog entry only proves the glass tower thinking without connection to reality, en.wikipedia.org/…/Marie_Antoinette ..

  88. susan says:

    @Joachim, too many out there are saying "oh edit pending.xml it will fix you up" when it doesn't.  This is a widely posted workaround.

    Personally the best advice I'm giving to folks is why?  Why install Sp1?  Do you need it?  Do you have a backup of your system before installing it?  What compelling reason do you need to install this service pack without having a backup of a system?  

  89. MaPaMa says:

    Once you get the error you are stuck, there is no "wait for the failure to happen". The suggested workaround just got the machine up again, and get the service pack to roll back. The suggestions of replacing the registry hives also leaves the system unstable AND forces a roll back. So which is better? neither solution gets the SP installed.

  90. Phil D. says:

    @Susan, I guess you are one of those that is still runing IE 6 too right?  You load a SP because it has new features and fixes.  I tested the SP on several machines in the office when it cames out and things were fine.  So when MS decided it was time to push the SP to a critical update, I said fine, I'll let my WSUS server push it out because I haven't had any problems in the office.  Little did I know (and anyone apparently) that if the user selected shutdown, which in Windows 7 automatically install updates that have been downloaded, that their machine would be rendered useless when they came in the next day.

  91. Phil D. says:

    "In most of the instances I have worked on with this, using the registry to rollback and then reinstalling the standalone package allows the service pack to install properly"

    It hasn't worked properly on any machine we have.  I'll try to get those logs to you today.

  92. susan says:

    @ Phil D. No I have Win7 here.  But I never install a Service pack soon after release.

    This has no new features for workstations that compells me to give the go ahead on WSUS for workstations.  

    Microsoft didn't push this, they stuck it on WSUS.  I as the admin decide when and it's exactly reasons like this that you never want to be in that first wave of dead bodies.  Wait.  Never be first.  Always have a backup.  The tortoise wins here, not the Hare.

  93. Dont test patches says:

    Why do admins never test the first rollout of patches from WSUS on test machines?

    Beats me.

  94. Phil D. says:

    Joscon, the problem JMPNY describes is the problem I was talking about.

  95. Christian Wimmer says:

    The official workarounds simpy don't work. Method 1 doesn't work since there are no restore points and method 2 doesn't word because that registry doesn't seem to exist in our case.

    We had 10 minutes to get the PCs working for a lecture and an exam. Doing an "unsupported" workaround suddenly didn't seem that bad….

  96. PA Bear, MS-MVP says:

    Just stumbled upon…

    Your computer may freeze or restart to a black screen that has a "0xc0000034" error message after you install Win7 SP1… (Revised 11 March 2011)

    support.microsoft.com/…/975484

  97. Drewfus says:

    In this context, does 'critical transaction' mean 'boot-critical', as opposed to Critical Update, or has it got something to do with Transactional NTFS? If it does mean 'boot-critical', then why can failure of the non-critical SP components to install via poqexec result in a no-boot situation? Also, why wouldn't running 'dism /offline:C:Windows /Cleanup-Image /revert-pendingActions' help here? Presumably the Pending Operations Queue Executor (poqexec) would run during any update that required a reboot, and therefore the DISM command would at least delete the SetupExecute value.

  98. anony.muos says:

    Again this just shows the unnecessary complexity of the Vista 7 servicing stack. Why and how was Xp servicing so simple and painless. You need to remove the complexity from this whole system of servicing and make it like Update.exe. i hope future Windows versions takes steps in this direction.

  99. Drewfus says:

    @someone: I don't necessarily agree with that. The extra complexity should buy you things, like flexability and reliability. That a service pack install can get a system into a state of limbo – neither here nor there and not booting either, doesn't necessarily mean that the servicing stack is too complex, it means Microsoft haven't (yet) leveraged the power that a more complex SS should buy. If the SP installation fails midway, reverting to the prior state should be as simple for the user as choosing an appropriate rollback option from the F8 menu, or running DISM with the /revert-pendingActions switch. If that command isn't 'powerful' enough, Microsoft should give us one that is – /revert-incompleteBuildUpdate – for example.

    Another more forgiving way of looking at this is that the OS has reached a point of complexity for which simpler technologies like update.exe are no longer appropriate. Massive feature lists necessarily mean massive complexity, including the servicing subsystem. The best way to appreciate the complexity and conflicting requirements that the architects of Windows have to deal with is to read about it, for example in books like Windows Internals. Not to discount the great info on this blog, but IMO a really thorough servicing stack white paper would be helpful (and apologies if it does).

  100. rahul says:

    @Drewfus: I agree with you completely on this. We seriously need something like /revert-incompleteBuildUpdate

    @Joscon: This is exactly the kind of situation that I was talking about when the miracle of 'Repair' was a tremondously a great answer.

    If we are able to slipstream SP1 into Window 7 DVD's easily and perform a repair (in-place atleast) that would be the answer for the incomplete SP install and the pending.xml hack used.. I am not sure if this possible.. We would still need a better resolution as this is only possible for smaller organizations for larger organizations like Phil's and JMPNY's still need better resolutions as this has impacted may be 100s of machines… As per Steve Ballmer at CES 2011 we have a Billion Windows 7 machines, and I agree that this has not been the case with all of the SP1 install attempts but even if it impacts 1-2% it is a huge impact……..!!!

  101. rahul says:

    Is there a way to slipstream SP1 on Win7/Vista ?

  102. gmflie says:

    Hi there,

    we also were forced to go the editing pending.xml file. If I have the original file can I "jump-start" the process again?

  103. Brian says:

    My network has 8 brand new PCs running Windows 7 64-bit in a SBS2008 environment with identical hardware that were set up exactly the same way with exactly the same software installs not more than a month ago.  4 of the 8 had this issue, the other 4 installed SP1 properly.  Bitlocker is not used on any of them.  SBS2008 forced the update due to the default group policy and I had half my users down during a critical time so the edit pending.xml option was used since it worked and I didn't know about the registry option.

    I find it troublesome that no restore points were available which would have made recovery much easier.  Shouldn't a service pack upgrade force the creation of a restore point?   And if, like Phil D says, the service pack removes the restore points, to me that's completely retarded since the purpose of a restore point in the first place is to recover after a bad software install.

  104. Rick says:

    Does that last statement imply that you can reproduce 0xC0000034 and that you're still not seeing restore point failures?  How are you reproducing 0xC0000034, assuming you are?  "Shutdown and install updates" for WSUS?  That seems to be the common denominator, though I wonder why it's not a widespread problem for home users, since they too can "Shutdown and install updates" (the update just comes to them in a different way). Perhaps the difference relates to WSUS serving the machines the monolithic SP1 .EXE patch (that's what I noticed WSUS has on our server anyway) versus WU just downloading for the home user the pieces of SP1 that it needs.

  105. Ben says:

    In my scenario, 7 machines updated fine, 4 machines produced C34 (all WSUS-deployed). Interestingly, the 4 machines are DELL OptiPlex 780 with an OEM Win7 Pro (German), the 7 that had no problems run manually installed retail versions (Win7 Pro German) on OptiPlex 760. Not sure if that helps. However, I had to fix them immediately so I went with the pending.xml-hack. Thus, here's another admin who would very much appreciate a simple way of fixing the follow-up problems outlined by you.

  106. princessid says:

    @Ben: you got me – I had one compay down – all with [DELL] OEM Win 7 Pro (WSUS – SBS 2008) – at another Company just those with an [DELL] OEM Win 7 Pro (also with WSUS – SBS 2008 network). All other PCs with a SystemBuilder Version are good (or got an error booting up with a "TimeOut" during the SP installation but came back up ok – reapplying the SP1 manuall worked fine)…. So maybe there is some kind of a difference…?

  107. princessid says:

    @joscon: personally I was also very dissapointed with what happend. 3 customers (2 of them with all clients down) called me wednesday morning. For a small business like mine this is a killer. All of them with a perfectly working SBS 2008 setup – up till now.

    The only thing I could find in the net (that helped btw), was to edit pending.xml. You can imagine that I can't tell a customer that I have to wait for Microsoft to come up with a solution.

    So what now – I'm unable to reapply SP1 on those PCs (I got passed the "missing objects" screen by manually copying the ATI stuff in filestore, but now the installer says that there are pending installs requiring a reboot – everytime).

  108. princessid says:

    OEM Win 7: I don't want to leave the impression that this is up to a dell installation, because I did a fresh installation with my normal Win7-Installation-Stick (that I use for installing all client PCs) on some of the PCs, just using the OEM key.

  109. Skiy says:

    My english is more approximate on sundays thant usual. Sorry for that.

    @flO: we also are small business and effectively, such event can be a killer.

    I can see that SBS2008 often appears (which includes WSUS) in comments.

    In our case, 80% of our clients uses SBS2008. We immediately stop deployment via WSUS.

    Also applied the pending.xml-patch on 10 PCs (9 Dell OEM Win7 64bits and one SONY VAIO OEM Win7 64bits) with obvious success.

    *None* of computers installed with retail versions of Win7 failed at SP1 installation.

    Hope this helps.

  110. Nick Whittome - MVP says:

    Jason,

    What I want to see happen is this.  (You already have all my logs etc).

    1.  A way to force Windows 7, in this state where a user has already edited the pending.xml file, to reinstall over the top with SP1 (even if that is a special download that Microsoft have to build and get up).

    2.  Never again for a SP to kill restore points.

    Cheers

    Nick

  111. Rick says:

    On the restore points, I'm curious, has anyone checked to see if they exist immediately after a successful SP1 install, or are they wiped only with a failed install?

  112. Skiy says:

    Actually, when passed the pending.xml hack , after rebooting, reserved space for restore points is 0.

  113. gmflie says:

    @joscon:

    with original pending.xml, same problem, after deleting the 2 Lines corresponding to 000.cdf-ms and left the checkpoint-entry in place. So far no issues with the installation, next thing I try is deinstallation of SP1 and reapplying it. Wish me luck.

  114. princessid says:

    @joscon: no space issues on the HDD. Today I had a lenovo-laptop at hand which I repaired with the latest recommendation you mentioned (remove the registry entry). Result: feels identical: can't install SP1 again for this failed installation…

  115. princessid says:

    @Nick: I do agree – I also want to have a SP1-"version" which I can somehow FORCE to install over those failed PCs…

  116. PGL says:

    I had 5 Dell Machines at my medical office fail to install the SP1 udpate via WSUS.  I tried the method 1 (support.microsoft.com/…/975484) resolution 1st and as others have said, there were no restore points available on any of the machines.  So I had to do method 2.  That did get me back to a bootable system, but I cannot install SP1 now (I tried manually installing it from the Technet download).  The machines think they already have SP1 installed and say something about missing components.  I'll follow this thread for a solution.  I would hate to do a complete reinstall on all 5 systems…

  117. Drewfus says:

    @Rahul: "We seriously need something like /revert-incompleteBuildUpdate"

    "If we are able to slipstream SP1 into Window 7 DVD's easily and perform a repair (in-place atleast) that would be the answer…"

    @Nick: "A way to force Windows 7 …to reinstall over the top with SP1"

    I'm looking at it like this – service pack problems could be divided into two categories:

    1. Installation (install fails and system mods and/or servicing hacks required to fix)

    2. Non-installation (functionality and compatibility issues – SP contents related)

    Now most of the problems we are seeing with SP1 are category 1 – installation failures due to things like automount disabled, non-active system partition, or whatever is requiring people to hack pending.xml (and what do we think is the root cause of that?). They are not due to the actual nature/contents of the SP – knowledge that should be exploitable.

    Ok, suppose we had a service pack that didn't actually do anything in terms of system updates. It simply installed components that were identical to those already in winsxs (except maybe the timestamps were slightly later). Now wouldn't many of these install failures still be occuring? If for example, your system partition wasn't set active, you would still see an error. However, you can easily back out using /revert-pendingActions (DISM), because no real updates are actually occuring – we just cancel the null upgrade and reboot.

    What i'm suggesting therefore, is that what's required is not a 'stronger' DISM switch (which probably can't work anyway, for the same reason that SPs can't be integrated offline), or a way to force-install SP1. What's needed instead is SP0. The idea is that you run SP0 across your network, and check that the install succeeded on all machines. If not, you block SP1, otherwise proceed with SP1 rollout. Waiting for SP1 to arrive before discoveing install problems is too late, and it's a bit like doing testing on production systems. You really want a null service pack to discover these problems before the real service pack install commences.

  118. Brian says:

    @joscon:  No drive space issues on the machines losing restore points – which is every machine with a failed SP1 update.  Each machine has at least 200GB free space.

  119. susan says:

    "@Nick: I understand what you're getting at but honestly, its not practical.  You could always parallel install the machines and move data using USMT to get them back up.  You could even do this with integrated SP1 media to save the SP1 headache piece."

    The reality of the SMB marketplace is that we may not have access to SP1 media.  We have OEM machines.  We may not have the rights to sp1 media, so there's no guarantee we won't be right back hitting this issue again.

    We need an inplace/over the top solution… please… understand we don't have all the media options at our disposal in SMB.

  120. Christian says:

    Hi Folks

    Got the same issue on some Machines. All different Hardware and differnet Software (except Win7 Enterprise x64) installed.

    Done the pending.xml-Hack because it was the First hint I found in the Web and it seemed to Work.

    But there is one difference: I did not delete teh Checkpoint, I commented out all thing behind the Checkpoint by simply adding "<!–" behind the Tag and "–>" at the end of the File.

    After I managed to reboot the machines, I allthough managed to remove the ServicePack simply using the Control-Panel.

    After that i reinstalled SP1 from Standalone-Package donwloaded at Technet.

    But now ther are 4 pending.xml-Files in WinSXS-Directory.

    pending – copy.xml (Backupcopy  I made of this File before)  Dated 03-09-2011

    pending.xml.01cbe2101a8fab27                                                  Dated 03-14-2011

    pending.xml.01cbe2066242a81e                                                Dated 03-09-2011

    pending.xml.bad                                                                               Dated 03-14-2011

    I think the users PC wasn't on since last Wednesday and Today the Error occured.

    But therefor the "real" pending.xlm-File now is missing, is it right that SP1 installed succesfully and there will be no more issues with an inconsistent State?

    Greets from Germany and sorry 'bout my horrible english!

  121. mike says:

    Joseph,

    Like many of the others, I had this problem on about 10 PC's, all running Windows 7 and 64 bit machines.  I had manually updated SP1 on some machines before it hit WSUS and those installed fine.  I had about 10 fail from the push.  I, too, did t he pending.xml workaround to get up and running.  

    I found KB975484 and like others method 1 is not a workable option, and I haven't tried method 2 as others have and it didn't work for them.  

    Will you please post the real fix for machines in this state as soon as you know?  I will keep following the comments on this blog post to seek a resolution.

    Mike

  122. 3guesses says:

    Note if you followed instructions to Pending xml you should have saved a copy as a back up that can be pulled back in no?

  123. rahul says:

    @Drewfus: I guess that we have chksur. It is not the SP0, but I presume it quintessentially does the same thing that you arr requesting. I'll leave joseph to confirm whether it is correct?

    Joseph: Still waiting for you to confirm your 'suspicions' and a newer post (with a KB likely).. on how to resolve this problem 🙂

  124. jtm says:

    I have SBS 2008 and I DID NOTapprove the install, yet WSUS deployed anyway. I have read other threads where this is the case. So far my machine made it ok. One at a remote location has the error 00000034 and can't get up. I think another at a remote location might have made it through (user sick today) and others are in question as they haven't shutdown yet. The machine that is down went through the "Shutdown and install updates" process, but so did mine which made it through.

    Does it seems that if you install from a "live" machine that the install goes properly (as opposed to "Install and shutdown")?

    Had seen some suggetion of that, but nothing concrete.

    Thanks to all trying to help.

    jtm

  125. rahul says:

    Joseph: This may be out of context here but have there been any issues with 2008 R2 – Hyper-V & Failover Clustering together- CSV or otherwise and any precautions that we need to take before moving to SP1?

  126. susan says:

    @jtm, the default settings for SBS 2008's (and SBS 2011) WSUS is to auto approve service packs for Workstations.

    You may not have personally approved it, but your WSUS settings did.  Go into the SBS console, to the security tab, change the client settings from high (which auto approves service packs) to medium (which does not).

  127. Kirk says:

    I suppose I'm in the queue waiting for information on fixing a system that was revived by editing the pending.xml file.

    Of the two systems that installed SP1 from WSUS before I removed its approval, one seems to have worked perfectly, and never produced any error messages.  The other had to be revived using the pending.xml edit.  These two system are as identical as possible given that they hav different main users.  All the hardware and drivers and applications are exactly the same.

    I am only speaking up here because, similar to Christian in his 14 Mar 2011 4:16 AM post, I did not quite follow the pending.xml edit instructions to the letter.  I only commented out the two lines immediately following the <Checkpoint/> line like this:

           <! DeleteFile path="SystemRootWinSxSFileMaps_0000000000000000.cdf-ms"/>

           <! MoveFile source="SystemRootWinSxSTempPendingRenamesa2226e58b7decb0144be00007000d017._0000000000000000.cdf-ms" destination="SystemRootWinSxSFileMaps_0000000000000000.cdf-ms"/>

    I left everything else intact, and on rebooting the SP1 install indicated that it was successful.  I did not try to reinstall because I was hoping to get more information before doing anything else, once the user's data was available.  He is also travelling this week, and had to have the laptop, so I can't do any further investigating until next week.

    I do have one more similarly identical system which has never tried the install.  I may leave it until more information is available.  I'll be watching this blog and whatever other sources I can find.

  128. rahul says:

    @ Kirk: I am hoping Jos has answers soon. We all are currently waiting.. and I stopped all SP1 plans.. Luckily I had planned it over the week-end of 12-13 March but after reading this. I decided i could wait..!! I thank my stars for another problem that creeped in last week.

    @ Jos: I am making an assumption that since this issue was prevalent since Vista SPs as per the initial KB article. There must have been some investigation at that time.. and some findings, analysis, etc. I am sure there must have been quite a few incidents due to which there was the KB.. Were there any resolutions to this problem at that time or was it let to die with time due to probable lesser impact? STILL AMAZED HOW DID THE DEV TEAM (OF THE LARGEST IT CORP.) LET THIS OFF AND LET IT CREEP INTO WINDOWS 7 ???

  129. mike says:

    Joseph,

    Any word yet for a fix on the machines with the pending.xml workaround?  Like others have said, I have several identical machines, and a handful updated perfectly, and others failed.  All of them that I have updated with SP1 from a DVD have worked fine.  I have 10 machines in the weird state, and would like to fix them as soon as possible.

    Mike

  130. Drewfus says:

    @Rahul: CheckSUR is going to fix (or at least, alert to) servicing/winsxs related errors, but it isn't going to fix some of the errors prevalent with SP1 installs, which as you know are often related to reading/writing to the system partition, BCD corruption, etc. So i don't think CheckSUR is another name for the imaginary SP0. If it were, you could have run CheckSUR before installing or allowing the SP update to proceed, and having seen errors, successfully predicted that the SP install would fail. Is that what happened?

    What i was thinking (smoking?) when i wrote about the non-upgrading service pack, is that ideally there would be a comprehensive means of determining if a system is service pack install ready. CheckSUR is one of the tools, but so is SFC, chkdsk, maybe memtest, and of course, there are the guidlines written by Joseph: blogs.technet.com/…/general-guidance-before-installing-service-pack-1-for-windows-7-and-windows-2008-r2.aspx Windows is a gigantic beast and no one tool can give you a completely comprehensive "go for service pack install" affirmation. However, the obvious goal would be to automate as much of this pre-service pack install testing as possible, including any "last minute" checks in the SP installer itself.

  131. mike says:

    @Joseph:

    Joseph, I'm sure you are, and I anxiously await a repair option.  My machines seem to be fine right now, I have just read where the pending.xml workaround might leave you in an unstable state and I would like to resolve that as quickly as possible.  I will quit asking and just keep an eye on your blog for the MS recommended repairs, as I assume you will post here when you have something to share.  I do appreciate all of the hard work.

    Mike

  132. Ellak Neniraas says:

    This is just a guess but have you considered language packs as a source of this error? We had +30 workstations affected with this problem. All had Finnish lang pack installed.

    When I started building a new Win7 image I noticed that if I install SP1 I can't install langpacks anymore. Lang pack installer says that OS is incompatible version. Removing SP1 allows langpacks to be installed again. This just made me wonder…

  133. rahul says:

    @ Drewfus: You sum it all: "Windows is a gigantic beast and no one tool can give you a completely comprehensive "go for service pack install" affirmation. "

    @All: Rewards are sweeter for the one who waits…!!

  134. princessid says:

    @joscon: got another Idea: I already said that nearly all my customers use SBS 2008 with WSUS. One customer was down completly (all DELL), the other one half, and another one with 0 problems (all PCs "selfmade").

    From the error-message I get after editing pending.xml and trying to install SP1 again, it seems to lead to the ATI driver (I have to manually add an inf-file and after that the SP1 just complains about pending updates).

    Coming to my first point: the Dell (customer1) are all on ATI, customer2 partly (didn't check 100%, but I know that the newer PCs are all on Nvidia), customer 3 isn't on ATI at all (all Nvidia Quadro).

    So to me it seems like all clients on ATI seem to have failed (in that WSUS-Domain environment), the others on Nvidia (and Intel?) are fine.

  135. ML49448 says:

    We have several machines down. All are HP Z400 workstations running 64-bit. All are running nVidia video cards. It does seem like the ones that failed are the ones that installed via "Shutdown and Install Updates"; all that installed manually are ok.

  136. R-D says:

    Don't think it is isolated to ATI, have had some custom systems with Intel integrated video go down as well. All in a SBS 2008 environment where SP1 came through WSUS automatically.

    Rather than using the pending.xml restore method, we manually restored the registry from the RegBack folder, this got the systems up and running.

    Afterwards Windows Installer didn't work and some other weird problems (cant print from Office 2010) were happening, all were resolved by running SFC multiple times on each system – running it multiple times was key as there is a limit it seems to how many files it can repair each run.

    Can you comment on whether or not this method leaves the system in an "unstable" state similar to the pending.xml method? The systems all "appear" fully functional afterwards.

  137. Leen Kleijwegt says:

    Is there a correlation between SBS 2008 – WSUS – W7 64 bits Workstation(s) and the installation of SP1?

    What surprises me:

    MS released a W7 SP1 Readiness Tool. Every time I used this tool all computers stated that they needed a Hotfix…

    Questions:

    Why did Microsoft release a W7 Sp1 Readiness Tool?

    Does SP1 also runs this Readiness Tool before installing the real Sp1?

    Is there an error in SBS 2008 WSUS? Why did the connected computers install SP1 while there is by default no rule installed to approve automatically this SP..

    To be clear:

    1 client with 4 computers, 2 identical HP minitowers and 2 identical HP Notebooks where 1 minitower has no problem at all with the Sp1 and the other 3 machines came to a not repairable standstill. All in a perfect working SBS 2008 environment till then…

    Who did something really wrong and what must/can I learn from this?

  138. dickfrey says:

    All our HP Z400 with Nvidia card failed the SP1. And because we have a WSUS, they ALL install by "Shutdown and Install Updates". This is the fatal combination: Wsus + shutdown and install.

    Probably the SP1 doesn't want a shutdown at the end, but only a reboot.

    But because the SP1 is available during a shutdown, THIS IS A BIG MISTAKE BY MICROSOFT.

    And when you have a lot of PC with engineers that are stopped, you can't say "now we wait for Microsoft solution". The pending.xml was the only solution at the moment….

  139. dickfrey says:

    The only PCs that had a successfull SP1 were uptaded WITHOUT a shutdown.

    And they are the same of the others with the problem.

    The problem was seen since 25 February. Why Microsoft didn't nothing ?

    Why they don't give a solution ? The ONLY solution was the pending.xml !

    At today there isn't a method for the C0000034 error, except reinstall the OS.

    But, as an users has tested, if you try the shutdown+SP1 in a clean OS, you will have the error.

  140. Joe says:

    @joscon;

    Out of 39 computers in my office, nine had SP1 installed manually and were fine. Tewenty nine installed SP1 at shutdown and failed with the c0000034 fatal error. One, the only non-Dell, installed at shutdown without trouble.

    I applied the pending.xml hack to all 29 failed Dells to return them to service. (I know, I know, but we had to stay in business.)

    I've tried reinstalling SP1 on some of these machines from DVD and so far I've had four successes and three that gave me the cannot continue due to missing components message. There's no obvious pattern to which ones work and which ones fail.

    Your help with this is greatly appreciated. If you're still collecting log files and such, I can probably scare some up.

  141. dickfrey says:

    @Joe

    "but we had to stay in business". It was the same for me. This is the reason of pending.xml method.

    How do you reinstall the SP1 ? The SP1 alone or did you use a MSDN DVD with the SP1 ?

    Thank you

  142. mike says:

    @Joseph,

    Joseph, I think I have said this already, but maybe not.  I have SBS 2008 and run WSUS defaults, or I did until this happened.  The default pushes out SP's to clients.  10 of about 15 clients gave the failure and I did the pending.xml workaround, so I'm still waiting on a permanent fix, and I hope it's note "reinstall", since about 1/2 of my users are remote and it would be a challenge to recoop their machines and rebuild them, but if I have to do that I will.

    Just curious, would booting into safe mode and trying to run the SP1 from there possibly work?

    You said you have not been able to reproduce, so does that mean MS is not close to finding a fix for this?  I truly don't want to rebuild a bunch of new machines, but I can if I have to, and install SP1 from DVD.

    Thanks again for the work on this issue.

    Mike  

  143. Joe says:

    @dickfrey & joscon

    I’m unable to uninstall SP1 on any ot my affected machines because it doesn’t appear in the list of updates in programs and features. So I just autorun a DVD created from the iso here:

    http://www.microsoft.com/…/details.aspx  Sometimes it works, sometimes no.

    The only noticeable anomaly in the cases where it works is that it never restarts the computer, regardless of whether I select auto restart in the setup or not. So I restart manually and I get the SP1 is now installed message on startup and SP1 now appears
    in the list of updates, along with about ten others, replacing the 70-80 that were there before.

  144. mike says:

    @Joe,

    I have tried to reinstall from DVD and I get an error about missing components, and it can't continue.  I tried this on 2-3 of my affected machines, all after applying the pending.xml workaround.  Do you have that workaround applied, and are still able to reinstall SP1 from DVD?

    Mike

  145. Joe says:

    @Mike

    I have a few machines that had the pending.xml workaround applied on which I've still been able to install SP1 from DVD. Others give me the missing component error.

  146. Drewfus says:

    Joseph,

    is it possible for you to hint or make an educated guess as to what is causing 0xC0000034 issues, at this stage? Is it any of the following?;

    0. Bit errors in WSUS server downloads caused by faulty network cards, cables, routers, switches, etc

    1. File corruption of the kind that chkdsk fixes, or non-fatal registry corruption

    2. Pre-existing servicing component errors or corruption that CheckSUR is designed to fix, and/or hardlink errors of the type that sfc.exe fixes

    3. Drivers or hardware related. Especially graphics card drivers, and faulty memory modules

    4. Non-default permissions on system-wide files & folders, services & registry keys

    5. Antimalware programs

    6. Malware, especially rootkits

    7. Non-supported image building practices. Images built with programs like "RT Se7en Lite" and "7Customizer".

    8. Other

  147. Drewfus says:

    Fair enough.

    Re fault reproduction, i wonder if anyone has seen this error more than once on the same machine, either after formatting C: drive and reinstalling using the product DVD, or using their OEM recovery images?

  148. ffwiki says:

    have somebody tested restoring the original pending.xml after the hack?

    I've read that someone had this idea but I didn't noticed somebody tried it…

    if yes – what happens???

  149. mike says:

    @ffwiki,

    I consider this to try to revert it back to the failing condition and apply the registry edits and see if that made a difference, however, when I booted back up in repair mode and when back through the steps to get to the folder where the pending.xml file was located, pending.xml was not there any longer.  The copy I had made was there, and there was a pendingXXXXXXXXXXX.xml file there, but pending.xml wasn't there.  I decided at  that point, not to rename the copy to pending.xml for fear I would render the machine unbootable.  

    It was working, albeit with the pending.xml workaround, and I didn't want to blow that up.  It was a production user machine, and I really didn't want to go through and rebuild from scratch to set it back up, as I do not want to do with any of my failing machines.

    @Joseph,

    I know I said I wasn't going to ask, but is any progress being made on  suitable fix?

    Mike

  150. mike says:

    @Joseph,

    cool.  Just checking.  I'm sure you guys are busy with this issue.  I guess overall, I'm not in a hurry, as my pending.xml workaround machines seem to overall be running fine.  Just worried about side effects starting to happen because they are in a presumed "unstable" overall state.

    Mike

  151. ryan says:

    Add me to the list of "pending.xml" fix users. Had a client that fixed it with that, after my suggestion of reverting the registry didn't work.

    Just a note, all my other machines (Dell Optiplex and Latitude) Win 7 Enterprise x86 machines seemed to have worked fine (They all have my custom win 7 image on them), but the OEM Sony x64 gave the error.

    All service packs were distributed via WSUS.

    Seems to me that x64 is the one to look out for.

  152. Nag'a' says:

    The KB article has been updated.

    support.microsoft.com/…/en-us

    Following "Method 1" on a machine currently at c*34 state should get your machine to SP1 without side effects.

  153. mike says:

    @Joseph,

    Good to hear of a possible fix for those still with the fatal error.  I would assume that is progress, and hope that we have a fix for those in the pending.xml state as well.  Joseph, thanks for all of your work on this and keeping us updated as we anxiously await a fix.

    Mike

  154. mike says:

    @Kelvin

    Thanks for posting your experiences.  I want to be sure I understand something you said.  On machines with the pending.xml fix, you were able to "fix" them by "upgrading" the install to a Win7 with SP1 integrated, and it successfully completed and the machine is now back to normal and working, with user settings/files and such?  Just want to be clear, as I might consider this if that is the case, especially for my remote users.

    Mike

  155. mike says:

    @badazzwinder

    thanks for the reply about what you have done.  I'm not going to try that yet myself, but it's good to know that mostly worked for you.  I had before this problem occurred and have since installed SP1 from a DVD of the ISO at least 25 times with no problems,   My problems came about because of WSUS, which is I think the case for most everyone.

    Mike

  156. dickfrey says:

    @Kelvin Aston

    But do you need another product code ? Or it remains the same ?

  157. mike says:

    @Kelvin,

    I have not tried this, I don't have a copy of Windows 7 with SP1 integrated.  Can I download it from Technet maybe?  IF that does seem to work, I am willing to try it if a fix from MS isn't forthcoming in the next little while.  My machines seem "stable" currently, so I'm willing to be a bit patient, but if this is a viable option, I will consider it. When you say "reactivate things like IE8, Search, Net3.5, etc…  what do you mean?  Also, if a new product key is needed, I'm not sure what to use there….  Can you and Badazzwinder provide a bit more detail of the steps involved to do this?  

    Mike

  158. mike says:

    @Jospeh,

    Joseph, can you comment on this being a possible real fix for machines with the pending.xml workaround?  Have you guys tested upgrading a machine with a version of Win7 and SP1 integrated, and would that be a proper way to get the machine in a working state?

    Mike

  159. Drewfus says:

    @joscon: "I havent seen an occurance yet where a reinstall hasnt allowed the service pack to work.  I'm sure it exists though."

    I'm a little surprised that no comments have been made about that statement, considering it suggests there is something wrong with the machines that see the error *prior to the SP install*, that is being eliminated by an OS reinstall. That counts out hardware, and WSUS issues, presuming that the post reinstall SP installs are via the downloadable exe and WSUS. So what is it? Surely someone is interested in the root cause, and not just their current predicament?

  160. mike says:

    @Joseph,

    Joseph, just for fun (yeah, right), I allowed a machine to try to receive the SP1 update via push, and sure enough it got the C0000054 error.  I grabbed the update KB975484, and executed Method 1 and that seemed to have worked wonderfully, and painlessly actually.

    So, can a machine with the Pending.xml workaround be put back in the state before the workaround and apply the KB975484 fix?  Part of the workaround involved creating a copy of the pending.xml file, but when I looked at a machine I had applied the pending.xml workaround to, it doesn't show me a pending.xml file in that winsxs directory.  I see the copy I made, and one that has pendingXXXXXXXX.  Could a person copy the pending.xml copy back to pending.xml, reboot into F8 and apply the KB975484 vbs script update?

    Mike

  161. mike says:

    "Reactivate things" means going into Turn Windows features on or off on the left side of Programs and Features in Control Panel. Features that you want on must have a check mark next to the feature. The failed SP1 update seems to have reset some of these features. I purchase Microsoft software through the Volume Licensing Program. This program provides one product key for each program or OS that needs a product key. This one key ,say for Windows 7 Pro, is used by all machines having this OS installed on it. Therefore, when I upgraded over the failed SP1 install, I just entered the same product key as the original install.

  162. mike says:

    The last post to you was from Russ aka Badazzwinger

    Russ

  163. mike says:

    Russ,

    OK.  I can go download via the Volume Licensing Program as well, I'm looking for Win7 with SP1 integrated, right?  Then, it's an "upgrade" install, and a few tweaks of features and all is well?

    If that is the case, I will probably try that on a machine the first of next week and if all goes well, then do the rest.  

    Thanks for the update.

    I still hope to hear from Joseph if there is a way to recreate the failed state after the pending.xml workaround and apply their fix.

    Mike

  164. mike says:

    You got it.

    Russ

  165. WinDose says:

    @mike

    if you apply Method1 from the KB on a machine alredy in an inconsistent state(by booting into F8), it wont work.

  166. susan says:

    Prob with using Technet/msdn media is that it's going to prob want a key and prob not like an OEM key.  You'll need to key with the Technet key breaking the licesning.

    I'm not sure you can install a slip over the top without it asking for the key code.

  167. rfxcasey says:

    Don't have anything technical to add here but my machine is borked and I need serious help fast. I tried using the script for editing the pending.xml file in the winsxs directory but it doesn't exist which has me a little worried. I am considering doing the reg key delete and recreate method. Sorry I haven't read through this entire thread, only half, but wanted to know what is the recommended fix now and what's the word from MS on getting this resolved. Just installed 7 for the first time the other day and its already broken. This is just terrible.

  168. rfxcasey says:

    Tired that I get an error stating Failure loading XML file d:Windowswinsxspending.xml and have no idea what to do now.

  169. rfxcasey says:

    Sorry to double post but just to add. I can navigate to the d:windowswinsxs directory manually but the only .xml file in there is called reboot.xml.

  170. rfxcasey says:

    @joscon do you have a link on how to do that?

  171. rfxcasey says:

    Tried this I typed in from D: directory which the recovery environment says is my system drive dism /image:d/cleanup-image /revertpendingactions and I am presented with an error:2 Unable to access the image. Any idea what I'm doing wrong?

  172. rfxcasey says:

    OK think I got it. I needed to put a : after the image:d making it image:d:. Seems to have worked though it gave me this cryptic message about the scratch drive might not be big enough but then it said operation completed successfully. I can actually get into a working 7 now so thanks so much for your help. Now the question is what about this outstanding service pack issue? Should I try to install the SP again, is there another way to install it or should I just wait till MS fixes the problem?

  173. mike says:

    @joscon,

    Joseph, just checking in to see if any progress is being made on a repair for those of us with the pending.xml workaround?  No huge hurry, just verifying that MS is still working on a a repair for those of us who used that workaround to get machines up and running.

    Thanks,

    Mike

  174. mike says:

    @Joscon,

    Joseph, I can't say that I am in a hurry, just requesting an update.  As far as I can tell, my machines with the pending.xml workaround are continuing to operate, so I'm willing to wait on a fix, just hadn't seen an update for a few days and wanted to ask. Thanks for responding.  

    Would you recommend rebuilding as the "fix", or do you think a suitable repair will be made available?

    Mike

  175. mike says:

    @joscon

    Joseph, no it makes sense, and as long as my machines are running OK (and I'm not getting any complaints currently), I will wait, as it might be a fairly straightforward fix once it is released (as the one from the update KB975484 – March 16th is – I applied it to one I broke on purpose, and it fixed it flawlessly).  

    Waiting for a repair, assuming one is forthcoming, is going to be better for me than rebuilding, as about 1/2 of my machines with the pending.xml workaround are remote, recouping those machines, and rebuilding them and redeploying them would be a significant hassle.  I would just as soon wait to see if a fix is coming.  Like I said, they all seem to be working OK for users, so I'm good for now.

    Thanks for your timely responses.  I know you guys are working on this and I appreciate the efforts.

    Mike

  176. dickfrey says:

    Like Mike, i will wait… Me too i have a lot of PC to rebuild, and now they are working with the pending.xml workaround.

  177. dickfrey says:

    Hi joscon,

    i've seen that i have a directori in C: called "749994f5b14d6c61f166df092a", and inside there' s a file called "spinstall.exe" of 590KB. Could be useful ?

  178. rahul says:

    I have waited for some info…. but seems that all roads are leading to rebuild..? Is there no resolution for this?

  179. mike says:

    @joscon

    Joseph, just checking in to ask if you guys are still working on a resolution for machines in the "torn" state.  Overall, I'm having no issues with those machines currently, just making sure I'm not waiting for no reason.  If you are still working on a resolution, let me know and I will continue to wait.

    Mike

  180. Spextor says:

    I also had some trouble with a couple of computers, some reinstalled and some in "torn state"

    I was just wondering if it's safe to install SP1 on Windows Server 2008 R2?

    If OK, I will do that from Windows Update as I have not had any trouble with Win7 Sp1 from Win update, just from WSUS.

    Has this error occured with Windwos Server 2008 SP1?

  181. mike says:

    @joscon

    Hey Joseph, I had asked this earlier but didn't see a response.  Can you somehow reverse the pending.xml workaround to make it fail with the C0000034, so the vbs script could be applied?  That works great.

    Mike

  182. mike says:

    @Joscon

    Joseph, perhaps you did and I missed it.  Anyway, that is what I was wondering.  I will go back to waiting to see if a fix is coming.  Thanks for the timely response.  I appreciate the effort on your part.

    Mike

  183. bibbler says:

    @joscon

    same problem with vista on customers notebook – but have sp2 installed but now have 62 updates saying they are installed then they are not – apart from this issue the pc is running very well

    bibbler

  184. hagman says:

    I have a question:

    We had several Windows 7 laptops showing the problem and used the pending.xml hack (not knowing anything better and there were no system restore points).

    All of these are currently travelling and hence not accessible for me to check.

    But: My WSUS reports that Win 7 SP1 64bit (KB976932) is installed on all of them.

    Can I *rely* on this report (i.e they are not in a state where they themselves think that SP1 is not installed or the like)?

  185. hagman says:

    Is a request for admin password for "issch" with every reboot another symptom of torn state?

    Is there still no suitable solution to un-tear the state for all of us pending.xml hackers?

  186. rfxcasey says:

    never did the pending.xml so I'm not in a torn state. Is it safe to download the service pack yet or is it still broken?

  187. mike says:

    @joscon

    Joseph,  Mike here, I was out of the office last week.  Any new news on repairing machines in the torn state?  Are you guys still working on that, or do I need to plan rebuilds?  Also, I did see where some have been able to "upgrade" with a technet version of Win7 with SP1 integrated, but that requires a new product key.  Are you guys providing one for clients in this state to repair these machines?  Just trying to catch up from being gone last week.  Thanks for any responses.

    Mike

  188. mike says:

    No they won't provide a key.  See if the OEM key will work.

  189. Reader says:

    @joscon: dead honest? i have been following this article for a long time now and gradually my suspicion that Microsof & you won't do anything to help those affected people mounts and mounts….

  190. rfxcasey says:

    @joscon Just installed the Win 7 service pack update which as noticeably changed since the last time I tried running it. All went well and I now have the SP1 update installed successfully.  Thanks for your help, you're a lifesaver.

  191. mike says:

    @rfxcasey;   Can you elaborate more?  Was your machine in a "torn" state, and if so, what were the steps you went through to run the update?

    @Joscon,

    Joseph, I do appreciate your responses.  I know this is difficult and people (including me at times) are impatient, especially since this is a MS created problem for us.  Anyhow, what are the detriments to leaving machines in the "torn" state if they are currently working?  As I have said previously, I have about 12 or so machines in this state, but no complaints from users and it everything seems to be "working".  Does it mean that no updates will be received by the machine?  I guess I'm thinking if there are no major detriments to the torn state, perhaps I can leisurely rebuild (if no fix is coming from MS) each machine over time, as many of mine are remote to me and I have to plan to go get them, leave a temporary replacement, repair/rebuild, and then return it.  

    Mike

  192. RoninV says:

    Though the 'pending.xml' edit does make the system usable for the enduser, it renders Windows Update unusable. Windows Update screen will appear with text (#elementModuleHeaderText#), and no options to use Windows Update. Absent a complete reinstall, this result might be a good reason not to do the 'pending.xml' edit. As an admin, the edit did come in handy when the bosses upstairs could not use their systems.

  193. mike says:

    @joscon; @Kelvin;

    Guys, if you test Kelvin's fix for the torn state and it is stable, will you post it here?

    Mike

  194. dickfrey says:

    @Kelvin

    If it does work, you need a MONUMENT builded by Microsoft !

  195. dickfrey says:

    Time ago i have putted the "atiilhag" folder missing.

    I remember i had another error after doing this.

    But i haven't setted the permissions, so i will do it.

    And then i will tell you something.

    Thanks

  196. Kristian says:

    @Kelvin

    I am indeed getting this:

    2011-04-13 14:10:57, Error                 CBS    SPI: Driver directory C:WindowsSystem32DriverStoreFileRepositoryatiilhag.inf_amd64_neutral_951c1812f542740a is missing

    I used SP1 DVD on a torn PC and checked the log after.

    Kristian

  197. dickfrey says:

    @Kelvin

    I have done your method and i can say that the points 7 and 8 are the second problem where i was stopped time ago.

    I have done the command on the point 8 (little bit different in something in the name, probably a ~ or similar, so i have copyed the right name from my CBS log, and the process has gone).

    After 3-4 reboot i have finally installed the SP1 from a standalone file of 903MB.

    It has worked for 45 minutes and at the last reboot it gives me an error, E_FAIL (0x80004005).

    But now in the history of Windows Update the SP1 results INSTALLED and COMPLETED.

    Before it was stopped at 33%.

    So i can tell that something you have resolved !

    And after the last reboot i have installed the upgrades released yesterday.

    I have about others 14 PCs to try, not now, but i will do in the future.

    Thanks again !

    @joscon

    If in Microsoft there is a place for 2 software engineers… don't search too much : "thiswoot" and "Kelvin Aston" are the two men that Microsoft needs.

  198. mike says:

    @Kelvin,

    Wow, great work!  I do want to clarify (I'm a bit slow I suppose).  You said you think the first two attempts to install SP1 from DVD aren't necessary, would that mean for someone looking to implement your fix you could start at step 7, and look at the cbs.log file and get the SP string, then move to step 8 and start from there?

    Mike

  199. dickfrey says:

    Excuse me joscon, but :

    – the problem born 25 days before the SP1 was released for the world

    – few days after, thiswoot gives a solution (not perfect, but a solution)

    – in the meantime, Microsoft has turned the head from outer side, why ? It was sure that the problem doesn't exist ?

    – when the horrible SP1 has released, a lot of PCs in the world are blocked !!!

    – The first solution (not working at 100%) from Microsoft is communicated 2-3 days after the worldwide trouble

    – so the ONLY solution to save our works comes from thiswoot, and all the workers can breathe

    – one week later, it arrives the second solution from Microsoft (and in the meantime ? Holidays for all? )

    – in Italy all the threads in Microsoft Answers that speak about thiswood's method ARE DELETED

    – and now after a month, another guy (not Microsoft) gives a solution for the thiswood's method

    Unfortunately for me i'm italian, but if i was a USA citizen probably I would have been the FIRST of the class action against Microsoft.

    And how Microsoft has managed the SP1 problem suggests a lot of thing for the future.

    For example: Why to buy a Windows Phone ? I can imagine the results of a similar SP1 on a smartphone…

    Excuse me again joscon, but i'm very furious for this SP1 and how Microsoft has managed it, leaving who work in troubles for a lot of days, and then "condemning" all the peoples that needed only WORK as soon as possible. And today, isn't clear why this error has come in our PCs.

    Best Regards

  200. mike says:

    @Kelvin,

    Thanks for the reply.  Will you post  a link here to your blog, I would like to follow it and see what success people are having?  I will want to try this myself, but I have to plan it out, all of my machines in the torn state are in production, so I likely have to wait until someone with the problem is out of the office for a day.

    Mike

  201. chucko says:

    @joscon

    IMHO, you seem to be missing the point that others are trying to make.

    It is poor practice IMHO to look at call volume as the primary indicator of how serious a problem is.  Most seasoned Techs don't call tech support, as it is usually a waste of time, expecially with MS.  Others may have found one of the fixes on the web and implimented them and once the PC came up they thought things were OK, as the PC shows SP1 being installed (but we know now that it is really not).

    You seem to be wanting to dismiss the problem because in your mind there are few PC's affected.  We all know that the numbers of PC's in a "torn state" is far higher than you think it is.

    More importantly the point needs to be emphasized that this is not our fault.  MS made the mistake here and needs to own up and fix it.

  202. chucko says:

    @Kelvin

    Is it possible that your posted fix is not for the same issue that we are discussing here?  The missing ATI files that your fix references I believe are part of another issue and are not related to the "torn state" issue caused byt the WSUS SP1 install.

    More info about the missing ATI files on this link:

    randombytesofstuff.blogspot.com/…/windows-7-plus-driver-sweeper-equals.html

  203. Andy Kinnard says:

    @joscon: I must strenuously disagree your assessment of Microsoft's mistakes here…yes, mistakes, not slower-than-needed response or something judged by perception or judgement.  Simply put, Microsoft should test ALL patches and service packs that will be pushes by WSUS in its default configuration using every available method of application INCLUDING the highly popular option to apply updates on shutdown.  In addition, this SP should NEVER have been published without its being marked exclusive.  For those two mistakes, there IS NO acceptible EXCUSE.

    I'm not describing extensive compatibility testing.  I'm describing basic due diligence.

    Beyond that, MS's continued explanation that they didn't detect the magnitude of the C34 problem because, somehow, the USERS failed (by not calling MS).  For those who need to use their computers every day to make a living, spending many minutes or hours on the phone with MS so they (or their IT) can help MS debug the problem (i.e., spending their company's or client's money to solve MS's mistake) is not a viable option.  Particularly since there was NO working work around on the first day, that meant that I didn't call MS that day because it would only delay getting affected PCs back in production.  After that, I found zero motivation (or compensation) for calling MS, furnishing them logs and reports, etc. for exactly the same reason: it does not help me, and I cannot get paid for it.

    MS is perfectly capable of testing this on their own: It requires no special hardware or third-party software…just a vanilla Win7 build that applies the updates from March (including SP1) via WSUS on shutdown.  Hey, MS!…stop pushing your work off onto me like I'm some sort of beta tester!  Grow up, take your lumps, accept responsibility, and fix both the problem AND the fallout (including fixing the torn installations resulting either from your work around or thiswoot's).  Nothing else is good enough, and I feel totally confident that my statements here represent those of the majority of users.

    I know it's hard for a company to do because there are no natural pathways for this, but, re-enable the corporate empathy process and try to understand how this impacted users.  Then reformulate your responses to be less condescending and denigrating (in content, not tone) about what users were forced to do to get back to business.  You all did this: users who used WSUS and applied updates on shutdown were doing what they have been told — what you advised.  Then, they did what they had to do to get back to work.  You all created all the conditions that led to that and could have prevented it all with a little due diligence.

  204. Andy Kinnard says:

    That's sure interesting because it seems to have happened uniformly in otherwise plain vanilla SBS networks with very new Dell Win7 Pro x64 PCs.  Every network for which I'm responsible (and some of my colleagues) and which is SBS 20xx with Win7x64 boxes had this issue.

  205. Andy Kinnard says:

    @joscon: We do appreciate this conduit of communication and your efforts here.  I'm sure you feel a bit beat up by the "kill the messenger" syndrome.  Never-the-less, some points made here are valid: call volume is an antiquated metric.  It may till be useful but doesn't measure well the impacted audience here: largely SBS networks.  It should be used in concert with other sampling data from Twitter, newsgroups, forums, etc.  I feel confident that, if there were a corporate initiative that required getting data about public impressions, that call volume would not be on the short list for measuring that data — the marketing folks would throw a fit until measures of internet traffic on the topic were included.

    When folks reply about things like, "This is a bigger problem than you think it is," and similar, I don't really think that they are trying to fill your mouth with their assumptions.  In other words, I don't think they have any idea what the numbers of affected users is or what you think it is.  I do think they are (perhaps ineptly) communicating a sense of abandonment and rejection as evidenced by a month passing with no clear solutions for the torn PCs (and far too much time passing before MS reissued the SP marked exclusive).  They are effectively saying that, if the response is congruent with the size of the impacted audience, then MS's measure of the size of the impact is inaccurate (because the response as far as solutions presented has been weak, as if responding to a relatively small issue).  I think they are saying, "This is a bigger problem than that suggested by Microsoft's response to it; therefore, since the response is supposed to be proportionate to the size of the problem, your all's estimate of the size of the problem must be in error: it's bigger than you think it is (whatever the raw numbers really are)".

    A month later, we really do feel dismissed…not necessarily by you, but by the company.  It's been a month. I've yet to check, but I wonder how those torn machines processed WSUS updates this month.  Any patching on those boxes would seem to me only to complicate matters, but that is the natural consequence of MS's taking so long to fully respond.  So, MS really needs to move on this one, even if it is just to say, "You're going to have to reinstall/repair-install/foobar-option," with some guidance on doing it in a way that is least disruptive and requires the least down time.  If that's the definitive fix; so be it.  Folks won't like it, but at least there would be definitive guidance.

    Again, we appreciate your efforts and so should Microsoft.  They should either pay you for this or put someone whose official job it is (to do this sort of thing) on the case.  It would be of no help to us or MS if you were to stop blogging here.  I'm sorry you end up being the recipient of all the anger and frustration.

  206. mike says:

    @joscon

    Joseph, I'm sure you are frustrated with the finger pointing, especially when it's directed at you, even though you are just trying to help, and I think most folks realize that.  I think the anger comes because we have been hampered by a problem, that at least from my point of view, looks like MS didn't do diligence in making sure this update was ready to be released.  Everyone I have seen appears to be in my position, a SBS network with the updates being auto-deployed by the default SBS configuration.  As I have mentioned before, I have about 40 computers in my network and I have 12 in this state, and this problem hit me on March 3rd..

    For me, when I had the 2nd machines reported with the C34 error, I went looking for an answer to resolve.  I, like many other, found the workaround from "thiswoot" and at the time I didn't know it would return unintended consequences.  I did not see a "fix" from MS until the March 16th update of KB975484.  The workaround did get machines back in a running state and that was important, especially since about 1/2 of my affected machines were at remote offices and not easily accessible for me, and I'm sure I would not have been able to wait a couple of weeks for a fix anyway.

    Anyway, with all of that said, I'm a bit surprised that we are now 6 weeks or more from this happening and MS still hasn't released a proper repair for it.  I will likely try kelvin's fix as soon as I can catch a user on vacation or out of the office for a day to give it a whirl, and hope others who try will post their experience here.  

    And I would hope you would share this with other MS engineers and see if it is valid and bless it as a repair for the torn state if that makes sense.

    Also, I am nothing but appreciative of you keeping this blog open and keeping us updated as to where MS is in the process of trying to find a repair for the torn state.  The update of the KB article on March 16 worked for me the two times I had the C34, I just wish it was out before I saw the workaround ,but that is the nature of things.  I, for one, have used your blog as the source of knowing what's going on with an issue I still have.  My machines are working with no complaints, but I know the torn state is not where i want to be long term, so I'm still anticipating a hopefully so to come repair option from MS.

    I am nothing but thankful for your attention and responses to my inquiries.  I appreciate all you are doing.

    Mike

  207. susan says:

    Andy?  Mike?  Ping me please if you would.

    susan-at-msmvps.com

  208. dickfrey says:

    I'm happy to see that a lot people thinks like me.

    The concept is : we are final users, not beta testers.

    I can understand joscon, he needs to defend his work and he's monitored by his bosses.

    But when a lot of people says the same thing, this is the truth, only this.

    Twitter, facebook, the forums. They are the new feedbacks for the big company.

    Because if one day we decide to buy another OS, we don't telephone to MS to tell him. Sure.

  209. Skiy says:

    It seems to be a dead-end.

    We are, for most of us, IT professionnals. When I passed the pending hack, I didn't like it.

    The idea is that we are professionnals, but the situation makes the (bad) solution, and when you look at several persons that can't do their jobs, you don't waste time to search the more appropriate solution. You do the faster. Sometimes, that's the job.

    We all appreciate what's Joseph doing here. That's not discussed.

    But in our case, I agree with Mike and Andy : MS have to solve the torn situation. We're installing everyday OS, Server OS, Applications, … Everyday we say MS is not perfect but we know them. Do we ?

  210. Kristian says:

    @ joscon

    "Currently the advice is to rebuild systems that have had the pending.xml workaround used on them."

    Any comments regarding Kelvin's fix? I'd like to hear more comments about it, from everyone. Is it working? Is it actually doing what we want here?

    Thanks,

    Kristian

  211. dickfrey says:

    You can see my comments upper.

    I have again 14 PCs to try, but not before 3 weeks…

  212. susan says:

    Tvarde – ping me and I'll get a case set up susan-at-msmvps.com

    Same with you dickfrey — I'll help you and you can help us all.

  213. Ray Avila says:

    Is the Microsoft KB script different from thiswoot’s solution?  Both seem to suggest modification of your pending.xml, and I’ve been avoiding that as advised by this blog.

    Joscon’s post on [13 Apr 2011 7:42 AM] seems to imply that the Microsoft KB script is a safe solution (aka Will not result in a torn state).

    To be explicit, I’m referring to the following Microsoft KB article:

    support.microsoft.com/default.aspx

  214. Drewfus says:

    @Andy Kinnard  "Simply put, Microsoft should test ALL patches and service packs that will be pushes by WSUS in its default configuration using every available method of application INCLUDING the highly popular option to apply updates on shutdown."

    To suggest that Microsoft doesn't do this is, at best, naive.

    @Andy Kinnard  "In addition, this SP should NEVER have been published without its being marked exclusive.  For those two mistakes, there IS NO acceptible EXCUSE."

    Great hindsight analysis. Now what do you suggest for these inexcusable mistakes? Public floggings?

    @joscon "–We have a small increase in calls at Microsoft on the first day that this hits but those quickly trail off"

    "All in all, I actually think that this is pretty good support for an issue that was mainly seen on the newsgroups and here.  Had everyone who had seen this C34 error called in the day they got the error, it would most likely have generated an even faster response for the script workaround to alleviate these issues.  I've said it before but I cant fix what I cant see.  When you dont call me for help, I dont "see" the problems you're having.  And when most of you dont call in, then the company as a whole cant see that.

    @chucko "It is poor practice IMHO to look at call volume as the primary indicator of how serious a problem is."

    A reasonable comment, followed soon after by …

    "You seem to be wanting to dismiss the problem because in your mind there are few PC's affected.  We all know that the numbers of PC's in a "torn state" is far higher than you think it is."

    chucko, is the second quote also your humble opinion? Who is 'we' a reference to? What are the true numbers? Why does joscon think they are less?

    @Andy Kinnard  "…MS's continued explanation that they didn't detect the magnitude of the C34 problem because, somehow, the USERS failed (by not calling MS)."

    joscon above indicates that Microsoft uses helpdesk call stats as a proxy for the extent of a problem and as a determinant for resource allocation for problem resolution. Andy, what evidence do you have that suggests MSFT "didn't detect the magnitude of the C34 problem", and that this implies end users have failed? Failed what? What is the extent of the problem? How has MSFT now managed to detect the true extent of the problem? How is it that you seem to know more about what is going on that Microsoft Support?

    "Then reformulate your responses to be less condescending and denigrating (in content, not tone) about what users were forced to do to get back to business."

    Who or what in general is this a reference to? It is poor form to be making a statements as negative in content *and* tone as you have without being in any way specific as to what has got under your skin.

    Maybe a few commenters on this blog should read, or re-read, the first post.

    blogs.technet.com/…/introduction.aspx

    to quote joscon: "…I am a Senior Support Escalation Engineer on the Windows CORE team.  What that fancy title means in laymans terms is that I am one of the guys on the other end of the phone when you call into Microsoft support."

    With respect to joscon – that is a position i envy – but remember people who your talking to here. This is not the head of the Windows division or the CEO your dumping your frustrations on. It's a guy who's doing his best to help, and educate, *in his own time*, even when problems like the subject of this post are obviously resulting in a lot of overtime. Please consider this when your commenting.

  215. Joscon (doesn't work for Microsoft) says:

    What a little liar…

  216. dickfrey says:

    What does it means this last post ?

  217. Humble fixer (MS "Owned" through a vendor) says:

    This is the way I've solved my problem:

    1. Press F8 on boot until the boot menu comes up. Choose the first option – Repair

    2. Choose Command prompt in the repair menu. When the cmd window appears type the folowing:

    d:

    cd d:windowssystem32config

    rename default default_old

    rename sam sam_old

    rename security security_old

    rename software software_old

    rename system system_old

    cd regback

    copy default d:windowssystem32config

    copy sam d:windowssystem32config

    copy security d:windowssystem32config

    copy software d:windowssystem32config

    copy system d:windowssystem32config

    exit

    Click RESTART

    3. Press F8 on boot until the boot menu comes up. Choose SAFE MODE

    4. in safe mode click Run and type services.msc and then ENTER. Find windows update in the list, dubleclick it and modify the Startup type to DISABLED

    5. Open My Computer, go to c:windows and delete the SOFTWARE DISTRIBUTION folder

    6. Restart computer