[John Hook]

arisme,

Did you upgrade or install any new drivers, Upgrade your BIOS, change any BIOS settings, change any Windows performance or accelleration settings or install any new software PRIOR to the point in time where you system started behaving this way?

I've had similar random freezing in Windows (NOT DOS) that usually occurred after I installed new drivers, tweaked some BIOS setting or changed some Windows display or performance properties setting.

As others have already stated here, improper CPU cooling (i.e. faulty CPU fan), faulty or overheating power supplies and other motherboard faults can also cause the random "lockups" you're experiencing. Overclocking, incorrect voltage and timing settings, BIOS shadowing and other settings in in your motherboards BIOS setup can also cause the symptoms you've described.

I've personally seen this most often with buggy video driver upgrades and/or video hardware acceleration settings cranked up too high.

If you haven't installed, upgraded or tweaked any of your Windows performance, display settings or motherboard BIOS settings prior to when this problem started - I'm thinking that somethings going faulty with your hardware (CPU / CPU FAN, Power Supply, etc.).

- John

[Advertisements]

Physical parameters look OK. I would also run a disk check by Maxtor's (Seagate) utility. Please use the following link and create a bootable CD or floppy disk and run a full scan of the disk.

Regarding Seamonkey, it looks like an obvious candidate, appearing so much in the Event log. Uninstall it temporarilly and use IE (preferably 6, not 7).

Regarding the USB error message: In Device Manager expand Universal Serial Bus Controllers, right-click each item and choose to un-install it. Reboot and let XP install it afresh.

Regarding Security errors (not red): could you write down an example?

Since the restart is not clearly linked to any program, we have to fumble after leads in the event viewer and hope for the best.

In general, I hope you don't mind me asking why your hard disk is divided into so many partitions. Usually it can lead to lot's of problems and waste of disk space because the fullest partition will affect the entire performance. In your case partition C is too full, having only 8% free space. That is not enough even for a decent de-fragmentation (for which you need at least 15%). Please see if you can free some space and run defrag.

My honest advice is that this computer could benefit a lot from clear format during which repartitioning will take place (for C partition I would leave at least 30 gigabyte). If you choose to do so please format to NTFS file system.

[The Skeptic]

Physical parameters look OK. I would also run a disk check by Maxtor's (Seagate) utility. Please use the following link and create a bootable CD or floppy disk and run a full scan of the disk.

Regarding Seamonkey, it looks like an obvious candidate, appearing so much in the Event log. Uninstall it temporarilly and use IE (preferably 6, not 7).

Regarding the USB error message: In Device Manager expand Universal Serial Bus Controllers, right-click each item and choose to un-install it. Reboot and let XP install it afresh.

Regarding Security errors (not red): could you write down an example?

Since the restart is not clearly linked to any program, we have to fumble after leads in the event viewer and hope for the best.

In general, I hope you don't mind me asking why your hard disk is divided into so many partitions. Usually it can lead to lot's of problems and waste of disk space because the fullest partition will affect the entire performance. In your case partition C is too full, having only 8% free space. That is not enough even for a decent de-fragmentation (for which you need at least 15%). Please see if you can free some space and run defrag.

My honest advice is that this computer could benefit a lot from clear format during which repartitioning will take place (for C partition I would leave at least 30 gigabyte). If you choose to do so please format to NTFS file system.

[arisme]

John,

Thanks for your thoughts.

The difficulty in pinning down the change that started the problem is that there is a time window spanning at least a couple of months, and probably more. ('change' here includes the possibility of hardware deterioration).

A new project was starting involving web services for the first time. As a newcomer to web services, I was scanning the net, downloading and exploring various commercial samples and trials, and developing software all at the same time. But in a studied and methodical sequence.

As far as memory serves, the first time it happened, nothing had changed on the machine and no new software had been installed for quite a while. So, I just shrugged my shoulders and carried on (without noting date and time - oh, for the gift of hindsight!). Some weeks later, second reboot. Power dips? Maybe. Some weeks later, another one. Stretching coincidence now.

I reinstalled some possible software suspects, and updated some utilities. Now it is clear that they were all innocents.

I've personally seen this most often with buggy video driver upgrades and/or video hardware acceleration settings cranked up too high.

Not touched or changed for nearly 4 years - the first 3 years being trouble free.

If you haven't installed, upgraded or tweaked any of your Windows performance, display settings or motherboard BIOS settings prior to when this problem started -

All were fine with original defaults - not touched.

I'm thinking that somethings going faulty with your hardware (CPU / CPU FAN, Power Supply, etc.).

My thoughts are veering in that direction, but I feel that I would like to eliminate drivers as exhaustively as possible first.

Aris

[arisme]

Regarding Security errors (not red): could you write down an example?

They all seem to be the same:-

Event ID: 615
"IPSec Services: IPSec Services failed to get the complete list of network interfaces on the machine. This can be a potential security hazard to the machine since some of the network interfaces may not get the protection as desired by the applied IPSec filters. Please run IPSec monitor snap-in to further diagnose the problem."

Possible causes: IE and IIS are blocked by my firewall, and need permission to connect to the net. I very, very rarely use either.

Regarding Seamonkey, it looks like an obvious candidate, appearing so much in the Event log.

I agree. In fact, when it first started being the active program at reboots, I reinstalled an earlier version which had run faultlessly for about a year prior to the reboot problem. Then, I recalled that the first 8 reboots had all occurred with another program, but which had been run much more frequently at the start of the reboots. So, frequency of running startups was determining which programs were likely reboot candidates.

One other clue - although I have broadband, I tend to disconnect unless there is a need to be connected. A fair number of the reboots have taken place when disconnected.

I will follow up your suggestions about the hard drive scan and USB driver reinstall, and report back.

In general, I hope you don't mind me asking why your hard disk is divided into so many partitions.

A mix of factors. To start with, I am a geek, and have been for more than a couple of decades. As a freelance software developer I have long running projects, and it is convenient to separate them into their own filing cabinets. Add to that a few peccadilloes, idiosyncracies and the way that my mind works best. Also, in the case of physical disaster, I can live with losing a partition, but the loss of all if there was only C: drive would be unbearable. The withdrawal symptoms from total loss would have me in a padded cell before you could blink.

So, C: drive is for Windows only (or it would be if there were not so much unthinking third party software that installs on C: without the courtesy of a user option). J: is pretty full with sources for downloaded utilities and packages. D: is for current copies of third party software.

When I ordered the machine in 2004, the original size spec allowed for Win98 SE to expand into the rest of C: at a gentle and leisured pace. What I did not anticipate was a project requiring XP Pro + IIS + .NET framework + miscellaneous other requisites, or how much bloatware that would amount to.

Last week I had 1.8Gb spare on C:. Then I read that I needed a 1Gb pagefile on C: in order for XP to do its BSOD diagnostic dumps. (The virtual memory pagefile has been on I: for a long time.)

I agree with you about the tightness of space - I normally aim to keep around 20%-25% free on each partition. I have another drive ready to add and reorganise. The reboot problem and current workload have got in the way.

> Please see if you can free some space and run defrag.

Will do, as soon as circumstances permit.

> My honest advice is that this computer could benefit a lot from
> clear format during which repartitioning will take place (for C
> partition I would leave at least 30 gigabyte).

We are as one on this - and one day I shall get there, time and opportunity permitting.

If you choose to do so please format to NTFS file system.

In my circumstances, I do not need more security or access control, as offered by NTFS. I have not experienced hard drive problems for more than a decade of 24/7 operations. (Of course, there may be a soft error lurking there as I write).

To be honest, I am somewhat nervous about the increasing number of "thou shall not" and "thou cannot" barriers that come with every step that I take in Bill Gates's direction. And I am not too keen on the one way street of FAT32 ==> NTFS.

What overwhelming advantage would there be in NTFS for me as a user to justify the switch? This is a sincere question made in the hope that there is a good technical argument that overrides my reservations.

---oOo---

In case I do not get another opportunity - what a wonderful forum this is, and I am full of admiration for the skills and experience that you and your colleagues volunteer for those in need.

Thanks guys. Very, very much appreciated indeed.

Aris

Edited by arisme, 17 April 2008 - 05:58 AM.

[The Skeptic]

If you are interested please have a look at this link. It's a comparison table between file systems. In my opinion there are at least 3 important parameters in which NTFS is superior to Fat 32: Max volume size, Disk space Economy and Fault Tolerance.

[happyrock]

One other clue - although I have broadband, I tend to disconnect unless there is a need to be connected. A fair number of the reboots have taken place when disconnected.

this is probably the reason for..Event ID: 615 "IPSec Services: IPSec Services failed to get the complete list of network interfaces on the machine.

are you connected to a router...if so... set it to stealth mode...don't disconnect to see if this error goes away

as for NTFS vs FAT32...go here...then you decide

Edited by happyrck, 17 April 2008 - 06:54 AM.

[arisme]

If you are interested please have a look at this link. It's a comparison table between file systems. In my opinion there are at least 3 important parameters in which NTFS is superior to Fat 32: Max volume size, Disk space Economy and Fault Tolerance.

Thanks. The one characteristic that caught my eye was 'Recoverability'. Is there a technical explanation of this anywhere?

Aris

[arisme]

Regarding the USB error message: In Device Manager expand Universal Serial Bus Controllers, right-click each item and choose to un-install it. Reboot and let XP install it afresh.

We have resistance. The uninstall operation becomes unresponsive.

I have two USB devices - a broadband modem, and an optical mouse.

The Device Manager's USB list has six items:-

SiS 7001 PCI to USB Open Host Controller
SiS 7001 PCI to USB Open Host Controller
SiS PCI to USB Enhanced Host Controller
USB Root Hub
USB Root Hub
USB Root Hub

Right clicking the top item and selecting the uninstall option brings up the Confirm Device Removal dialog. Clicking OK changes the cursor to the hourglass, and the broadband modem lights extinguish. Five minutes later, nothing more has happened, and the Task Manager tells me that the uninstall has become unresponsive.

The same happens with the second item; the mouse is deactivated and the software becomes unresponsive.

Instead of trying to uninstall the devices, should I be uninstalling the drivers?

Aris

[The Skeptic]

The method I proposed removes everything that is related to the device and should work well. The idea is to force reinstallation by rebooting and letting Windows reinstall all the drivers available in windows. The fact that it doesn't work is not a good sign and I worry about carrying on with that. Please reboot and go to the same place in device manager, double click on each item and see in General > Device Status if the device is working properly. If it does, leave it at that. If it doesn't try to update it or uninstall and reboot (in the Driver section).

[arisme]

The fact that it doesn't work is not a good sign and I worry about carrying on with that.

I agree - somewhere in the bowels of the system something has gotten out of line. Could it be a registry key that is either corrupted or has been left in an incorrect state?

Please reboot and go to the same place in device manager, double click on each item and see in General > Device Status if the device is working properly. If it does, leave it at that.

Done - all reported as working properly. However, I have never been able to trust those reports entirely after past ecperiences where they have turned out to be inaccurate.
In this case, both devices do work.

The method I proposed removes everything that is related to the device and should work well. The idea is to force reinstallation by rebooting and letting Windows reinstall all the drivers available in windows.

Is there another way to induce the wholesale driver reinstall?

Aris

[arisme]

Skeptic,

I have just had a very interesting experience.

Having asked the question about inducing a forced reinstall of all drivers, I decided to see if the net could provide an answer. I stumbled on this:MS KB article

It refers to a utility shipped with XP (and some other Win flavours) called Verifier. This verifies installed drivers in a variety of ways.

I ran it (command line) and chose the option to verify every installed driver. Clicking Finish in the Verifier window brought up a dialog to reboot the computer and "implement the changes". I am guessing that Verifier stores the user options, and then applies them at boot time.

So we went into a warm reboot. After the BIOS screen etc, and about the time that signs of life would normally start appearing, a BSOD came up.

The text included

A device driver attempting to corrupt the system has been caught.
The faulty driver on the kernel stack must be replaced with a working version.

But there was no identification of the actual driver.

The stop code was 0x000000C4, which has this elsewhere on the MSDN site:

Bug Check 0xC4 DRIVER_VERIFIER_DETECTED_VIOLATION, defined as 0xC4, is the general bug check code for fatal errors found by Driver Verifier. The following parameters are those passed to KeBugCheckEx and displayed on the blue screen.

The parameters were 0x60, 0x00, 0x2000, 0x01, which probably will be of no immediate help.

I hit the reset button to restart the computer out of the BSOD, and the first screen up after the BIOS activity was the DOS-like option screen for restarting in various modes. The choices included Last Known Good, and Normal Windows. I chose Last Known Good, in case Normal still had the Verifier scheduled for another run (which might have trapped me in a cycle of Verifier BSODs).

This incident very nicely confirms the 0xD1 stops experienced with the random reboots. So, the current diagnosis of the random reboot problem is definitely pointing at a device driver - but which one?

Well, what we know is that reboots do not happen in idle periods, with or without program windows on the screen. As far as I can tell, with two exceptions, every single reboot has happened when a program has been fired up and before its display has started to develop fully (in some cases while the splash screen has still been on show).

The two exceptions are with program windows fully open, and the reboot has happened when some program activity has been initiated.

This gives us one common characteristic for all the reboots - triggered by program activity.
Another characteristic is that all the programs suffering from reboots have been run many, many times without problems. The implication of this is that the faulty driver is called only occasionally, or is called every time but only fails occasionally. Or is it called by one of the many Windows background tasks, and falls over in a coincidence of particular foreground circumstances?

Hmmm....much food for thought.

If the Verifier utility is new to you, would you take a look at the testing options? There might be something in the other choices and settings that will enable your expertise to suggest a way forward from here.

Aris

[techpro5238]

Hello Arisme,

Random restarts are usually either caused by a RAM corruption, or overheating. I believe overheating was cut off when you said most randomly, so I would have to point this to a RAM problem.

Can you please use Memtest86+ and run it for about 2 hours. Then post the results as to whether it has gone bad, or it is another issue.

Regards,
Techpro5238

[arisme]

Can you please use Memtest86+ and run it for about 2 hours.

I have already run Microsoft's Windows Memory Diagnostic for several cycles without any reported problems.

If you think that Memtest86+ would find errors that are missed by the MS tool, let me know and I'll run Memtest86+ when a time slot permits.

Thanks for your interest.

Aris

[techpro5238]

Memtest is of better quality. It also tests the CPU Cache, and other places on the computer.

For all we know this could be related to your CPU. Memtest will allow us to figure that out.

Go ahead and run the test for about 2 hours.

Regards,
Techpro5238

[The Skeptic]

Memtest is not better then Microsoft's tool but it could be a good idea to run a longer test, even throughout the night. The two programs do the same things.

Generally speaking technicians do not try to pinpoint the source of the problem because that is practically impossible with our knowledge and tools. What we try to do is to locate the general environment in which the problem evolved and find a wide-range remedy that will cover many options in this environment, solving the problem in the process. We never know what file or code line were corrupted. This is done (as you surely know) by software developers.

I honestly can't say where the problem is. The more we discuss it the more I agree that the source is in the software. There are a number of option which I could think of:

1: Run System File Checker. Insert the xp cd into the drive, click Start > Run. Type sfc /scannow and confirm. Let the process run to the end, at which the dialog box will just disappear. This process will compare installed system files to the ones on the disk and will replace them if necessary. It can be very useful if a corrupt window file cause the problem, but will not fix any application or non-window file.

2: XP repair. This is a process that I hardly ever use because many times you end up with more problems then you start with. If you wish to learn about or perform repair, you can use the link in my signature to a G2G guide.