Frequent unresponsiveness, 100% hardware disk usage, 0% software usage

[SpringHalo]

Every 10-40 minutes, the application I'm currently working in [whither it be skype, L4D2, or firefox] becomes unresponsive, and after a minute or two, the window turns white, notifying me that it's not responding. This unresponsiveness quickly spreads to other open processes, rendering them frozen as well. I can still use the mouse, and it appears rainmeter still responds [updates, showing barely any CPU usage].

Steps I've taken to try and resolve this:
Using only one main program, still does it
turning the computer off and on again
Memtest is clean
chkdisk passes
MBAM/MSE scans return nothing
defragged
no memory overflows that I can see

While troubleshooting, I noticed that during the 'fit of unresponsiveness' the disk usage light on my tower is solid, but when I looked at the disk usage in the resource monitor, no processes whatsoever were using it [screenshot related]

If any more specifications are needed, I'll be happy to elaborate.


Thanks!

 

[Lemur]

It could be indexing. What if you turn off indexing just as an experiment?

 

[SpringHalo]

I've disabled it long ago. I use everything now.

 

[Lemur]

Try running msconfig, then diagnostic startup. See if you have any problems.

You can also look at startup in msconfig, and temporarily uncheck these.

 

[SpringHalo]

Doing so seems to have fixed the problem, though there wasn't much to do to check if it would happen again, since there's nothing to do in diagnostic mode.

 

[MilesAhead]

If it flares up again you may look to see if any of these are enabled:

WMP sharing media files over the network

WMP logging the location of every media file on your system in case you might want to add it to the library at some future date(I'm not kidding.)

Network shares set up to run in offline mode(local copy is cached so the system keeps checking the remote copy for changes.)

Some application monitoring a large folder for changes, such as the Windows folder.

Also you may try disabling or opting out of the Customer Experience Program. The system gathers all kinds of stats(I forget if this is in W7. I know I had to turn it off in Vista.)

Vista my drive churned something fierce. Windows Seven the problem was diminished but it still calmed down after some tuning.

 

[SpringHalo]

I never use WMP, so I would doubt that it's starting any shenanigans right now. As for something indexing or searching, firefox could be the culprit, as a while ago I created a ramdisk for it to use as a cache, and now the ramdisk is gone. I looked back to change the cache location to default, but the bit I changed seems to be gone. That's the only thing I can think of off the top of my head.

 

[MilesAhead]

That's the thing about WMP. It was being started by a service in the background to index the media files to share over the network. Even though I never clicked on it and set other players as default player. I finally got the settings to the point I can use WMP if I want. Version 12 has PGS and XSub subtitle support. So on occasion I'll take a quick view with it to check subs.

I used to run Firefox portable out of a Ramdisk totally. Not just the cache. Best way is to discard the Ramdisk contents on system shutdown. Use XCopy to copy the Firefox portable folder contents recursively into the Ramdisk at logon(you can find the batch file if you search the web.)

Only real downside is you have to start FF portable from the HD folder if you want to store any settings changes or install AddOns. But it loads quickly even if the HD is busy from the Ramdisk. Kind of fun, but then I moved to Chromium. Even with multiple processes I start Chromium with a blank page so it loads fast if HD is busy.

Some people just load the browser and leave it open. But that would be cheating.

 

[SpringHalo]

Do you know which service it was that was causing the problem?

Edit: My problem isn't solved yet. When I said it seems to have fixed it, I meant that the problem didn't occur in diagnostic mode, from what I could tell.

 

[cluberti]

If you're willing to bugcheck your box "for science", you can set it to be crashed manually via the keyboard. Once you set the appropriate CrashOnCtrlScroll registry value for either a PS/2 or USB keyboard (depending on what's attached to your computer) and configuring your machine for a complete memory dump, you should reboot for the changes to take effect. Then, the next time your machine goes into such a state, you can simply hold down the right-hand CTRL key on your keyboard and press the SCROLL LOCK key twice to bugcheck the machine (on purpose, of course). This will generate a rather large memory.dmp file in %windir%, but that can be zipped up using 7zip or WinRAR to a much smaller size, at which point the zipped file can be uploaded somewhere for us to download and analyze the hang condition.

 

[SpringHalo]

I read about that, but I can't seem to find a scroll lock button on my keyboard. Would it be preferable to have as few programs open as possible at the time?

 

[SpringHalo]

Double post

 

[fseal]

I had this same problem some time ago. It was on a new SSD install though.

The issue turned out to be a problem with the disk drivers "NCQ" (Native Command Queing).

I had to go to the hardware manager and to the SATA port for the drive itself and on the advanced options I turned off NCQ and it never happened again.

Only some drivers expose that option though, but it's worth taking a look...

 

[SpringHalo]

Sounds like it could help. I'll give it a try when I get back to my computer.

 

[MilesAhead]

Do you know which service it was that was causing the problem?

Edit: My problem isn't solved yet. When I said it seems to have fixed it, I meant that the problem didn't occur in diagnostic mode, from what I could tell.

I have Windows Media Player Network Sharing Service disabled. Unfortunately I don't keep a log of every system change. I know I turned that service off. The rest of the tweaks I believe are in the WMP settings themselves.

Also if I remember correctly when you set up the network with the radio buttons for the various settings, it wants to have WMP share media. After making the settings I wanted, I went back into the configuration on several occasions to make sure my settings "stuck." For example if you log out, then log back on, or reboot, take a look at the config to make sure it didn't go back to what it wants.

 

[SpringHalo]

I've found something peculiar. When I enabled the memory dump option and commenced dumping when the freeze happened again, the computer just stayed in bluescreen for about 2 and a half hours without doing anything. At the time of the dump I had about 1.7GB of ram used, and 2GB of SWAP. It shouldn't take that long to dump, should it?

 

[cluberti]

No, it really shouldn't. It should also note that it's committing memory to the dump file with a counter at the bottom, so it would seem that perhaps it wasn't deciding to dump at all if you didn't see a count up to 100...

 

[SpringHalo]

There was no counter, just something saying 'preparing memory for dump' and 'preparing HDD for being dumpee' or something of that nature.

Edit: it said collecting data for crash dump
initializing disk for crash dump

With no progress bar or counter underneath.

 

[cluberti]

Hrm - sounds like the hard drive hardware has gone out from underneath. Once Windows gets to a blue screen, it has very little information left running, but it does have a very minimal driver for hard disk access to dump all data to pagefile.sys, and then reboot to rename it memory.dmp in the \Windows directory (by default). If the drive running the OS has gone south at that point and is no longer physically available, though, it won't have enough logic available at that point to do anything but wait with what you see there. This could still be a driver issue (as part of it is still used at that point), but it could also be a hardware issue with the underlying disk subsystem or the disk itself. If it works in diagnostics mode, though, it would seem more a driver issue than a hardware one. If you don't see any issues in the event log regarding your disks, then while it is still worth checking the SMART status of the drives to be sure, I'd put money on this being a disk controller driver issue.

 

[SpringHalo]

How would I go about checking this SMART activity? I'm currently running an error scan on the hard drive itself, but I'm not sure how to check the driver

EDIT: Downloading updated motherboard drivers, just to check.