Bug in Windows 7 RTM

[Antman]

Also, I thought rck01 was the Kennedy person for whom so many of you possess disdain. As far the referenced Kennedy goes, I have only read two web pieces written by him. Both were lucid and cogent. And something about no pants.

Back OT - I am pursuing this anomoly for a single reason: there is nothing on the interweb regarding it's cause or solution. An unsourced citation regarding something someone said is irrelevant and amateur. To this point, MD's treatment remains the singular attempt at professional analysis that I have seen. The treatment, however, remains speculative, to a degree, and does not address the discrepancy between dllhost, explorer and chkdsk memory consumption. The discrepancies are prima facie evidence that the analysis may not be on target.

I do not know of what value it would be, other than for pride - imagine that SevenForums were able to identify the specific cause and proposed solution before the hotfix is released.

 

[Mark]

Disk Check on all 3 drives and no problems on Build 7600.16384 x64, G is the external drive and B is my second internal Drive.

 

[Antman]

Disk Check on all 3 drives and no problems on Build 7600.16384 x64...

Can you show the processes page? How much mem was allocated to the task?

Do you have an "updated" chipset driver?

 

[Mark]

I'm using Nvidia Nforce Win 7 chipset Drivers and on G drive it was using 500MB, Primary used about 400MB and External used about 250MB.

 

[Antman]

I'm using Nvidia Nforce Win 7 chipset Drivers...

I am not asking which chipset driver you are using, rather which version. There is an unsubstantiated report that legacy (pre-RTM) mfgr's chipset drivers are at fault for the observed phenomenon. Is your chipset driver dated after the release date of your OS install?

Also, the reported issue does not seem to affect the sysvol. What is the file count on your G and external drives? Low file counts are consistent with (relatively) low mem usage in other reports.

 

[Mark]

Drivers were pre-RTM Win 7 drivers, version 15.35 signed 8th of May.

 

[mickey megabyte]

what's happened to ant's and grim's avatars???

 

[Muad Dib]

I guess I missed all the Kennedy stuff?

Anyway - turned off UAC - Disk Checking via Tools is now running as Explorer AND mem usage hit 3.22GB within seconds. File count on the target drive is constant from yesterday.

Cancelled scan and Explorer DID NOT release memory. I was able to close that instance of Explorer with a right click end procees. It closed immediately with no visible affect.

FYI...

I was able to duplicate the results, with Explorer - not CHKDSK, as described above. My first reaction is - THAT is a bug. I attempted to shake it up by starting numerous secondary apps as I had in my original tests with CHKDSK. No luck.

While in the Resource Monitor-Memory tab I right clicked the hung Explorer process (after cancelling the repair) and selected "Analyze Wait States". It reported that one of the threads in the process was waiting for a Network I/O to complete....(?)

CHKDSK, DLLHOST, Explorer .... Command Window, Properties - Tools Menu....hmmmm......

 

[Tews]

what's happened to ant's and grim's avatars???

It seems that Antman has morphed into a monkey named MoJo and Grim is doing his happy dance!

:roflmao:

 

[Antman]

Drivers were pre-RTM Win 7 drivers, version 15.35 signed 8th of May.

Interesting. Now I am curious and will do some Googling - are AMD systems affected? Are nVidia chipsets affected? What is the commonality?

Also, I still do not have enough data to fully grasp what is different about your result. a) I do not have enough non-RAID drives to run multiple instances of chkdsk (et al) to duplicate your 3 drive test. I do not have 2 drives that I can unmount (you can bet that I will revisit my drive configuration though). This is important only because you seem to be running multiple instances of chkdsk in your posted result.

I am not willing to run chkdsk /r on my RAID volumes. Please advise, anyone, if I am reticent without reason.

In order to corelate file count to mem usage per drives, I need to know the mem usage per instance of chkdsk.

Forward: chkdsk will now mean chkdsk.exe, dllhost.exe or explorer.exe unless otherwise noted.

 

[zigzag3143]

It seems that Antman has morphed into a monkey named MoJo and Grim is doing his happy dance!

:roflmao:

Maybe its just rule 15 and the admins having some fun

btw they got mine too

K

 

[Zidane24]

Interesting. Now I am curious and will do some Googling - are AMD systems affected? Are nVidia chipsets affected? What is the commonality?

Also, I still do not have enough data to fully grasp what is different about your result. a) I do not have enough non-RAID drives to run multiple instances of chkdsk (et al) to duplicate your 3 drive test. I do not have 2 drives that I can unmount (you can bet that I will revisit my drive configuration though). This is important only because you seem to be running multiple instances of chkdsk in your posted result.

I am not willing to run chkdsk /r on my RAID volumes. Please advise, anyone, if I am reticent without reason.

In order to corelate file count to mem usage per drives, I need to know the mem usage per instance of chkdsk.

Forward: chkdsk will now mean chkdsk.exe, dllhost.exe or explorer.exe unless otherwise noted.

Nice new avatar Antman...MooooJoooo-JoJo

 

[Zidane24]

Maybe its just rule 15 and the admins having some fun

btw they got mine too

K

Ken snip the post above....2222 posts

 

[Antman]

what's happened to ant's and grim's avatars???

DJG revealed my secret identity in the x86 versus x64 fight. Cloaking momentarily offline. I need Petrossa's flux capacitor to reset the timeline. My cockroachbots are on recon.

 

[Mark]

I just ran it again and checked the process Tab and explorer went from 123MB to 954MB and it was constant in it's ascent so I guess if there were fore files then it may have gone considerably higher, also you will notice in the first screen I posted of the performance Tab, the memory usage did go back to normal immediately after completion of each test.

 

[Antman]

...it was constant in it's ascent so I guess if there were fore files then it may have gone considerably higher

...also you will notice in the first screen I posted of the performance Tab, the memory usage did go back to normal immediately after completion of each test.

I misinterpreted your screenshot.

But, I do notice that you are not duplicating the test. Both check boxes should be selected to emulate chkdsk /r.

Apples and oranges. Can you test again?

 

[Mark]

This is taking a bit longer to complete now, 30 minutes in and....

 

[Antman]

Now you have seen what the buzz is about. Interesting that you get ~1GB, ~2GB and ~3GB results. I have not cross-referenced these to your file counts, I am assuming they corelate. If they do, your results are consistent with the MD's initial analysis.

I am glad that I did not bother with the Google re: AMD/nVidia.

While in the Resource Monitor-Memory tab I right clicked the hung Explorer process (after cancelling the repair) and selected "Analyze Wait States". It reported that one of the threads in the process was waiting for a Network I/O to complete....(?)

Muad Dib, MD, will always be "The Doctor" in my mind as a result of this examination. Once I get an acronym in my head, I can't shake it out.

 

[Muad Dib]

I duplicated my original test to confirm that CHKDSK /R executed from a command window does relenquish its memory usage when aborted by closing the command window or via the Task Manager - Process Window.

 

[LFB]

Access is denied...

Can't run here... what are the commands? CHKDSK <C:> /r

I have two HDs let's see if this crash my system...