Solved hangs overnight

[rhodesengr]

Machine is a Dell precision T5500. The system hangs while unattended. It always hangs if left overnight. It sometimes hangs if left for shorter periods like over lunch. I ran sfc.exe which found some corrupted files but this did not fix the problem. I have run the machine in safemode for over a week with no hangs. I run memtest for 30 cycles with no errors. I tried running a registry cleaner. This did not fix the hang proble. jcgriff2.zip file is attached. I was not able to get the perfmon file. When i run perfmon /report, the window opens and it keeps saying collecting data. It never finishes and creats the report. Any help would be appreciated.


- x86 (32-bit) or x64 ? 64 bit
- the original installed OS on the system? W7 Enterprise
- an OEM or full retail version? OEM
- OEM = came pre-installed on system
- Full Retail = you purchased it from retailer

- What is the age of system (hardware)? two months
- What is the age of OS installation (have you re-installed the OS?) have not reinstalled

 

[Jonathan_King]

Hello,

I suspect this is being caused by a hardware problem; however, software is also definitely worth looking into, as the problem does not occur in Safe Mode.

I see your Broadcom Ethernet drivers are rather old:

5/28/2009 6:57:35 PM   C:\Windows\system32\DRIVERS\b57nd60a.sys

Install these updated ones: Broadcom.com - Ethernet NIC NetXtreme Server Driver Downloads

Could you run driver verifier? See this guide: http://www.sevenforums.com/tutorials/101379-driver-verifier-enable-disable.html

Hopefully that will flag any bad third-party drivers, resulting in a BSOD. If that happens, upload another jcgriff2 report.

If we diagnose this as a hardware problem, Dell should be able to replace the bad part, or the whole system, as it is within the 1-year base warranty period that Dell provides.

 

[rhodesengr]

before your reply, I updated both the broadcom driver and the nvidia drive. I set up the verifier as explained in the linked post but I am unclear what happens with that. Am I supposed to get a report somehow? Do I do something else with it?

I guess I'll have to wait and see if the system hangs over the weekend since the new drivers were installed..

 

[Jonathan_King]

Driver Verifier runs in the background, silently testing each driver for stability problems. If it comes across a bad driver, the computer with have a BSOD, which we can then analyze for clues.

After you enable Driver Verifier, keep using the computer normally (but saving your work often!). If, after enabling driver verifier, your computer doesn't crash for 48 hours, go ahead and disable it.

 

[rhodesengr]

BSOD over the weelend

well the new drivers did not seem to fix the problem. Sometime over the weekend it went to a BSOD saying "hardware failure....system halted". It would not even boot at first but I went through a couple of repair cycles. The repair program seemed to say it could not fix the problem but after a couple of tries it did boot normally. I have attached a new jcgriff2. Still not sure if its hardware or software. Any help would be appreciated. Oh, I did not try to re-run sfc /scannow yet. Thought I would post the new jcgriff file first.

 

[Jonathan_King]

It seems this is being caused by bad hardware, honestly.

Have you run Prime95 yet? If not, run the "Blend" and "Small FFTs" tests, as described here: http://www.sevenforums.com/tutorials/100352-cpu-stress-test-prime95.html

Also, have you enabled Driver Verifier yet? Please do so, if you haven't already.

Final question: have you contacted your higher-ups about this issue yet? Has any diagnostic work been done by anyone else, or did you post here first thing?

 

[rhodesengr]

I enabled driver verifier back on Friday per your suggestion. Is there a separate log or dump file I should post from this? It is still enabled. I have not run Prime95 yet but I will now. I have had a help request to our internal IT people for over three weeks but they have not been able to get to me yet so I have been trying to fix it on my own.

Can you say what makes you think its hardware? I don't know where to look in the jcgrif file.

 

[rhodesengr]

now running prime95

First just wanted to say thanks for your help. Its very frustrating having a brand new PC that does not work and i can't get help from my own organization. To some degree I don't mind trying to fix it because I learn a lot in the process such as these various tools you suggest.

I am now running prime95. I have a couple of questions. I have started it in small fft mode. how long before I should try blend mode? How can I tell if I get errors?

Also, I can add that as far as I know, my internal IT guy is planning on "re-imaging" the hard drive. If it still fails after that, they will declare its a hardware problem.They probably will not spend a lot of time testing and doing diagnostics.

 

[Jonathan_King]

The only log we could find from Driver Verifier would be a crash dump file, which does not exist. If Verifier finds an error, it will cause a BSOD; the dump file from which will be picked up by the jcgriff2 program. Apparently you have had no BSODs yet, which makes it sound like drivers are not at fault here.

What makes me think it's related to hardware? Well for one, you have very few errors in your event logs, which I would expect to see plenty of if software is the cause. You have a few errors, yes, but I can't imagine they are at all related to your real problem, the hanging.

Now that I know Driver Verifier has been running, without causing a BSOD, I can use that to blame hardware.

That "hardware error...system halted" message you posted earlier is a good indication hardware is the cause.

The fact that it took several tries to get into Windows and the fact that Safe Mode works fine are somewhat confusing to me. Problems that don't exist in safe mode are usually caused by software.

Also, if this is caused by hardware, why doesn't it happen while you are using it actively? Perhaps that could be explained by a power management feature, such as Intel Speedstep or something else.

As a first step, try updating the BIOS. I see it is dated 4/12/2010; not old, but not the most recent either. The BIOS can be found here: Drivers & Downloads

I don't suppose you have a floppy drive on that thing; if so, that would be the easiest method of updating the BIOS. If not, you'll have to use Linux. Now be warned, I have never updated a BIOS this way, so I can only give my first suggestion. If that doesn't work, have a play around.

If you use Linux, I suggest that you create an Ubuntu Live CD: Download | Ubuntu

Boot from the CD, select "Try Ubuntu...", and then run the BIOS file from there.

I strongly recommend you get clearance from your boss before you do this. Flashing the BIOS is one of the riskier procedures you can do, as a failed update can leave the computer as a brick on your desk, requiring a replacement BIOS chip.

See if updating the BIOS alone does the trick; if not, post back here, and we'll go from there.

 

[rhodesengr]

not getting memory.dmp or minidump

Also, I do not see either minidump or memory.dmp files in my windows directory. I thought I set things up to get the minidump.

 

[Jonathan_King]

You won't get a dump unless you have a Blue Screen of Death (BSOD).

 

[rhodesengr]

The following are the instructions for the bios flasher. It seems to say I can just download the file and run it from windows. I don't have a floppy drive (CD only) and it would be nice to avoid making a linux CD.

------------------------------------------------------------------


Download file to Desktop


The file T5500A07.EXE is using the Universal (Windows/ DOS) format and is designed to be directly

executed from Windows environment.

1. Click Download now.
   The File Download window appears.
   
2. Click the Save button.
   The Save As window appears.
   
3. Select Desktop using the drop-down menu to the right of Save in.
   
4. Click the Save button.
   A progress meter moves as the download is completed.
   
5. If necessary, click X in the upper right-hand corner of the Download Progress Window to
   
   close it upon completion.

NOTE: Shut down all other applications prior to running the BIOS update utility. The update utility

reboots your system in order to get the new BIOS loaded.

 

[Jonathan_King]

Go for the Windows method. I was thrown off by these instructions:

This file format consists of a BIOS executable file. To use it, download the file and copy it to a DOS-bootable diskette. With the diskette in the floppy drive, reboot the system and run the program.

Trust the ones you posted above the others. I was surprised to see the former directions anyway, as Dell has a Windows-based installer for their other systems.

 

[rhodesengr]

BIOS updated then some boot problems

The BIOS updater seemed to work fine from windows. I saw an "update successful" message before the first reboot. This boot failed. I screwed around with booting from my repair disk and after a few tries the system booted from the HD without the repair CD in the drive. I didn't do anything in particular to fix the boot problem. I tried the repair function once which reported that it didn't work. I noticed that during the boots, i would see a blue screen with a bunch of white text flash briefly and then it would start to boot again. The blue screen did not last long enough for me to read.

Anyway, it does seem to boot now from the HD with the new bios. I have attached a new jcgriff file from the current boot. I still don't see any minidump files.

One of my problems is that I do not have a Windows installation disk. If i did, I would be re-installing Windows at this point. They use some special disk image that has a bunch of special software they use here to monitor our systems and push updates out.

Maybe you can see something from the latest jcgriff file. BTW, the PRIME95 blend test run without errors. Now I'll have to see if the system hangs overnight again.

 

[Jonathan_King]

The BIOS was successfully updated:

BIOS Version/Date    Dell Inc. A07, 11/30/2010

The first-time boot problem may be normal, I don't know. Sometimes those BIOS updates need a couple of reboots before everything is applied and back to normal again.

You shouldn't be seeing any minidump files, because you have gotten no BSODs,

 

[rhodesengr]

cautious optimism

system survived overnight for the first time in a while so perhaps the bios update fixed the main problem. I tried running "perfmon /report" and it still does not stop data collecting and generate a report. Is that a fixable issue? Anything else i should do to see if the system is actually stable?

I appreciate all your help.

 

[Jonathan_King]

I honestly don't know about that perfmon problem. Perhaps a system files check might help, or a repair install. The latter is not possible, because as you said, you don't have a Windows installation DVD.

To do a system files check, open a command prompt with Administrative powers, and enter sfc /scannow.

Don't stress the perfmon file though, unless it's an issue you really want to get resolved. I rarely do anything with that, to be honest.

The best way I can think of to determine the problem is gone is just normal usage; leave the computer running at night, as you normally do, and watch to see if the problem happens again.

 

[Jonathan_King]

Received via PM:

Its been crash free since I flashed the BIOS. Thanks again for your help.