Chkdsk bug in Windows 7 RTM Build 7600.16385/16399


Recommended Posts

I just ran a chkdsk /r myself (/r is for locating bad sectors by the way - something most people won't be doing often) and stopped when chkdsk was using 5GB of RAM. I am seeing the same results even on a chkdsk /f (which fixes problems - chkdsk on its own will just scan and not fix).

However when doing that, the scan is over quick enough for the memory usage not to get too heavy, although I did see it hit 500MB before the scan finished. I'm testing on build 7229 as I don't have the RTM build yet.

I've always thought of it as a old fashioned 'boot to floppy - run in dos' kind of thing!

Well you wouldn't be able to scan the system volume but you might scan another volume, no doubt about it. And being as you're going to have to boot to a Windows PE type environment to run this normally, I'm sure the bug would be present in that environment too..?

Confirmed. And yes this should be considered a showstopper. chkdsk killing a machine is not cool as there are real world uses for that tool.

I agree, especially for the tech support industry, but also system administrators.

Remember how the 16385 wasn't released as final RTM because one or two OEM drivers weren't up to par. That's how little it takes. This is much more serious than flawed drivers for a few particular kinds of hardware in my opinion, as it's not discriminating on hardware.

Edited by Jugalator
Just tried it, uses a lot of ram, computer stays responsive, ram is released when chkdsk finished. I'm not sure this is a bug at all to be honest, and even if it is, definitely not a show stopper.

Maybe the OP got a BSOD because he has a bad RAM module which normally isn't used, do a memtest.

Yes, it's a bit hard to tell whether a crash is due to faulty hardware or something else, like (perhaps unrelated) software starting to misbehave when the available RAM suddenly plunges. I've seen system instability as a result from radical memory loss in the past on Windows, so I think it's anyone's guess at this point. If you run a clean Windows 7 with not many apps running and start chkdsk, I can imagine there being less problems seen.

It seems some people chicken out when the ram usage hits a very high amount (especially those with more than 4GB of ram)

Any brave souls who lets CHKDSK use more than 4GB of ram and managed to finish the whole process without a BSOD?

It not occurring on Vista might not mean it is a bug in Win 7. It MIGHT be a feature..but no one can affirm to that.

But one thing is for sure, that the system stays responsive even with large amount of system resources being allocated to CHKDSK.

Where is Brandon?

There are more than one bug in the RTM, I have found little bugs or glitches already...

Just chosen to ignore them as they do not harm and forgot bout them, but Im sure some exist..

Anyone get Windows Live glitch when person signs into windows live somtimes..

Also whats with this rate and improve your pc crap...one most money making features of windows..Got scores all 7.3 and over and hard drive transfer rate 5.9 total 5.9 suggesting go buy a SSD drive what crap..

This isn't that big if none of you spotted it until you force ran chkdsk/r.

However, imagine this could be nastier if it affects 2008 R2 also (which it likely will)

Hm.. I've got a VM of Server 2008 R2 installed.. I'll go check it now!

Such a huge flaw, oh my gosh stop the children!

It's not a huge problem, it has not affected anyone up until now where a single person got a bsod probably due to bad ram.

chkdsk won't run on an activeely locked partition anyway and asks to be run at a reboot schedule but how often is the command used as said above?

Hardly ever,

Once again, a minor bug is blown out of proportion Neowin style.

so many people using software illegally. :rolleyes:

Actually I haven't cracked my install, I am waiting for a serial from MSDN. Not illegal :D

yes there is !!!

Number place in the world pirates

Contribute something to the thread and kindly quit trolling please

Will they make a new build and release it on 6 Aug? or will it be delayed?

No new build, probably just a hotfix

I don't think it is a bug. Maybe it is by design to speed up the whole checking process. I ran it on my laptop with RTM x86 7600.16385 and left it alone running for about 10 minutes on an external WD 500GB drive. Sure, the RAM usage increased quickly from around 700 MB to 1,701,732 KB in matter of seconds but it stayed there without increasing or decreasing the usage for a few minutes and my computer (again, a laptop) was very responsive.

This would not be by design, it could seriously cripple some systems with lower amounts of ram

I wonder if it only used 600k of ram because I checked a small partition of about 100GB ..

The partition I checked on my laptop was only 78GB

Uh oh - getting it on Windows Server 2008 R2 as well...

Wow, I wonder how far back it stretches :/

anybody that running seven rtm is running a illegal os it not the 6th of aug yet

I find this hilarious because I saw a comment posted by you on a Windows 7 torrent on Demonoid where you linked to Neowin in it. So don't try saying that wasn't you :laugh:

Hardly a showstopper IMO :/ serious? may be, showstopper? not. If only OP got a BSOD(bad RAM?) then it is more likely to be a feature where it uses available RAM as it should. Does it significantly slow down overall system?

Actually I haven't cracked my install, I am waiting for a serial from MSDN. Not illegal :D

As for illeagle software!! YES IT IS ! as there are no official legitmate downloads out yet! you must have it from a torrent ! therefore illeagle

SO is this happening on all those *LEAKED* builds...with the chinese altered files? SO you guys are sending an unknown third party all kinds of data from your computer, when you think your chkdsk is bugged? hahahaha... :)

chkdsk Who does this anyway?

I don't think this is a bug. When chkdsk is performed things needs to be loaded into memory.

I read some discussion on this from another site. The reasons why it seems to be a bug are that Vista does not do the same thing (I ran the same command earlier on a 500GB volume and sat at about 16MB RAM usage for the chkdsk process). Additionally some folks suggested it might be some performance enhancement but people have tested the Vista and the Windows 7 chkdsk commands side by side / against the same volume and have not been able to identify an increase in performance - just a huge increase in memory utilisation.

I simply don't believe this is by design. On a server you could quite legitimately take out one volume for some maintenance whilst serving up other volumes - to then basically consume almost all available memory would seriously degrade server performance. How could that ever be 'by design' ?

SO is this happening on all those *LEAKED* builds...with the chinese altered files? SO you guys are sending an unknown third party all kinds of data from your computer, when you think your chkdsk is bugged? hahahaha... :)

Just to be clear, I've only tested this on Windows Server 2008 R2 release candidate - not on any dodgy downloaded / hacked version of Windows 7!

Yeah, I think it's pretty safe to say that if this IS a real issue, Microsoft will patch it with Windows Update.

Odd, I think I read somewhere that people are already getting Windows Updates coming through, I've only had Windows Defender updates so far.

I'm running 7 x64 RTM right now (7600.16385) and have not run into anything CLOSE to a showstopper bug (in fact, the only bugs I'm running into are driver and third-party application bugs, and darn few of even those; the only *driver*-related bug I have run into since RC, involving the Sound Blaster X-Fi and hardware-compatible clones from companies such as Auzentech and ASUS, has a fix available from ASUS). The only fixes/patches (other than MSE/WD definition updates) are the usual Patch Tuesday fixes (for vulnerabilities that Windows 7 or applications therein shared with older versions of Windows, mostly in terms of IE).

I actually liked Vista (started with the early betas, stayed with it through launch and the first two Service Packs); however, 7 is a lot less painful from a user standpoint than Vista was. (Build 6801, the first widespread leak of Windows 7 that I ran across in x64 trim, was the first build I ran outside of a VM (after running the 32-bit build in a VM for a week); except for some later side-by-side tests, I've stayed with Windows 7 either in a dual-boot with Vista Ultimate x64 or, since over a month after the CPP kicked off, running 7 x64 as sole Windows operating system.)

Plus you only need the /r flag if your drive has bad sectors and you want to attempt to recover readable data from those sectors anyway.

Do people not chkdsk /? first to find out what the flags do..

Just checked this on my main pc. Got up to 1.9gb after doing a standard scan (No /r option). Excessively high in my book.

Lots of folks think that almost ANY RAM usage is excessively high, despite RAM prices unheard-of in PC history. (While older DDR and PC133 SDRAM prices are heading up, both DDR2 and DDR3 prices are continuing to slip south; I can buy DDR2 PC-6400 (200 MHz quad-pumped) for $15/GB retail today. DDR3-1333 DDR3 is now within ten percent ($16.50/GB or 6 GB for $100) of that.)

While Windows (especially 7) can be run on a RAM-starvation diet, it's no more healthy to do that to a computer than to do that to yourself. (Yes; I'm the same guy that ran Windows 7 x64 in a gigabyte of RAM.)

If you have 2 GB of RAM or less (even if you're running XP in 32-bit), you should upgrade RAM before upgrading *anything else* except hard drive space (and if you have used less than half your primary hard drive, before even that). That is true of 64-bit Windows to an even larger degree (especially if you have Office 2010 on your radar, and you should; the 64-bit version is that good).

At current prices, unless you're running an older desktop or laptop that takes only DDR, staying below 2 GB is silly at best, and stupid at worst IMHO.

[As for illeagle software!! YES IT IS ! as there are no official legitmate downloads out yet! you must have it from a torrent ! therefore illeagle

I happen to take issue with your inference. I contacted Steven Sinofsky this evening from my legit copy of the RTM 7600.16399. Their working on the issue. For those that are ignorant to this tools usage. don't bother posting unless you know what your talking about. :huh:

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.