Some problems, some symptoms, one clue

[wallypiper]

For history of my machine, you can read this long thread.

As you can probably guess from the end of it, my system ended up stable. Since April I have had zero problems. Not a single BSOD or other error of any kind except for a mouse failure that turned out to be exactly what it sounds like, a failure of the hardware itself.

A few days ago, while I was away on business, my wife called to say the computer "was on that screen with all the numbers and at the bottom it says 'hard disk failure, insert system disk and press enter'" or roughly that.

I had her power the entire box down and restart and she got the same error so I told her I couldn't walk her through the next steps over the phone and to just power it back down and leave it till I got home (she's got a netbook she can use when this box is not available). Of course, this gave me heartburn. I routinely back up important data from my HD but a genuine HD failure is still a major pain.

However, the next morning she sent me an email saying she had turned the system back on and it booted and was working normally. Hmmmm. OK. Thanks.

So this morning, I got a BSOD "page fault in a non paged area" while just looking at email.

So I ran the Seagate test program (just the short version) with no errors. Then I ran Memtest86+ v4.00 through a pass with no errors. Then I ran Prime95 and got a fatal error on core 1 during test 5. "Rounding error was 0.5, expected less than 0.4." The high temp recorded by cortemp was only 69, which is actually the lowest high temp of the four cores.

So what does it mean? Glitch? Heat? (It's not hot in the room at all.) Power supply? Karma?

[Hqxriven]

Re: Some problems, some symptoms, one clue

Seagate SATA Barracuda 7200.11 1.5TB

Firmware??

Seagate barracuda 7200.11 drives failing - The Inquirer

Seagate Barracuda 7200.11 Problems

[wallypiper]

Re: Some problems, some symptoms, one clue

PS - a new run has now made it through test 6 on all four cores with no errors, about 10 minutes. I know that's a very short test but it passed the point where it generated and error the first time. CoreTemp is reporting 75-68-71-72 with recorded highs of 77-70-73-74. I can hear the cooling fan stepping up and back occasionally. The test is still running but I'm going to stop it for now because I have other things I need to use the computer for right now.

[wallypiper]

Re: Some problems, some symptoms, one clue

hmmmm. I'll do a little research on the Barracuda issue and see if I can update the firmware on my drive.

I've already considered getting another drive and doing RAID. Is that hard to do? Tricky in any way?

[wallypiper]

Re: Some problems, some symptoms, one clue

Seagate says no firmware update is available for my serial number.

[Chike]

Re: Some problems, some symptoms, one clue

Core temperatures are high, especially that you are running stock I assume, are they the same when you got the machine stable?
How about the ram, does it feel hot to touch?
Are all case fans working properly?
What is the ambiant temperature?

EDIT: what is "system" temperature, as reported by EasyTune or other software?

[kick]

Re: Some problems, some symptoms, one clue

You have had some history..
I didnt read thru all the original posts,just the early and later pages.
Id be inclined to go along with Chilkes ideas on temperature.
Standard settings shouldnt give temps that high.
On the fans.do u have any others aside from the cpu and gx card and the one in the psu?.
Since the pc has been running awhile it might be time to do a "crud search" inside the case.
Im sure your house is very clean .but if there is any dust in the air the pc seems to attract it like a magnet and that will reduce the cooling efficiency.

Did loading and saving "optimised defaults" in bios never work for you?
I didnt read thru all the thread so maybe u did that before.
Sometimes when you get a bsod the os will say hdd failure and it might be just corrupted data from the system lockup, but it needs to check the hdd ,or usb drive (if you use readyboost) anyway.

Other than that,check that all old drivers have been uninstalled ,nvidia ones especially used to have a problem with leaving certain bits in the registry.
Do the dxcheck to make sure gx is ok.

[wallypiper]

Re: Some problems, some symptoms, one clue

Yes, those core temps are about what I saw back when the machine seemed to become rock solid last spring. All the fans appear to be working.
At the moment, ET6 says the system temp is 41, but the machine is not doing much.

[wallypiper]

Re: Some problems, some symptoms, one clue

Its hard to answer the optimized defaults question briefly. Basically, when I load optimized defaults, I don't seem to get the BIOS settings that people expected to see. In any case, there are some settings that are not optimized defaults at this point. The very end of the long thread has some minor tweaks that were done at the end and a some earlier pages have images of all the BIOS screens. I have had zero, I mean absolutely zero, errors, crashes or even things that go bump in the night since the end of that thread with the exception of the mouse failure. It appeared at the end there that there were problems with the Nvidia driver but that seems to have been resolved. I have many fans. I am not using the fans that came with the Corsair memory as one of the three had problems and I decided that I would just try running without them rather than have a sitcky one pulling on the power supply trying to run. I have seen cases, not with PCs but with other fan cooled electronics, where a sticking fan would pull the voltage down far enough to create other errors.

Anyway, I have the stock CPU cooler, an exhaust fan on the back of the case, a supply fan on the side of the case and the front of the case and the fans built into the psu and graphics card.

I did find some cat hair built up on the side fan screen which I have removed. Right now, with nothing going on but the browser, Core Temp says the cpu is at 43/40/40/42 and ET6 says the system temp is 40 and the cpu temp is 33. ET6 says the cpu fan is at around 1440 rpm. It doesn't report any other fan speeds. I believe all the other fans are just full speed.

[Hqxriven]

Re: Some problems, some symptoms, one clue

Seagate 1.5TB Drives Randomly Freezing

http://technologyexpert.blogspot.com/2008/11/seagate-releases-firmware-fix-for.html

http://techreport.com/articles.x/16221

[wallypiper]

Re: Some problems, some symptoms, one clue

Interesting stuff on the Seagate issue, especially the fact that they aren't posting the update on their website which is, as you can see in my post above, what I encountered. There was no explanation on the website, by the way, regarding contacting them via some other method. They have a utility that checks based on the serial number of the drive and it simply reported that no update for the drive was available. At this point, I'm not certain that the HDD failure incident was legit. Maybe some other problem created it as kick suggested above. It is not the first time this box has reported an HDD failure. Back when I was trying to get it stable, they happened several times but the drive has worked flawlessly since then.

Meanwhile, I ran Prime95 for several hours yesterday while I as out and all four workers eventually stopped due to errors. The high temps recorded by Core Temp were still in the same range - 70-75 C.

I looked into better cooling for the processor a while back and gave up after finding that there seem to be very few upgrades that will fit in my case. It's narrow with only about 160mm between the surface of the processor and the opposite side of the case.

I picked up a new HD yesterday to install as a backup drive (I have an external USB drive but the speed and the fact that I don't and don't want to keep it attached all the time make it a very occasional backup solution). It's a Hitachi Deskstar 7K1000.C ($60 at Microcenter!) l'll install that and get some regular back up set up (I routinely back up critical info like personal financial data, but not the mass of documents, pics, music, software downloads etc. although all are backed up at long intervals). While I have the box opened, I'll give it a thorough cleaning inside and then maybe run Prime95 again to see if that made any difference.

Any other suggestions?

[Chike]

Re: Some problems, some symptoms, one clue

I'd still look into lowering the temps.
What is the room temperature?
What thermal paste do you use, and have you removed and installed the CPU HS without applying new paste?
Try running Prime95 for 30 minutes with the case open and room fan blowing over it and see if there a significant change in core temps and others.

[Lsdmeasap]

Re: Some problems, some symptoms, one clue

I think you can update the firmware so that both drives match or have the latest, serial number checker only checks for firmware related issues then advises (I believe).

I think you can update it no matter what the advisor says, you just need to find the firmware you want to use.

[wallypiper]

Re: Some problems, some symptoms, one clue

Well, things have really gone down the toilet today.

After I installed the second drive, the computer would not boot. It did not even list the first drive in the bios screen.
But it was seeing the new drive but it wasn't formatted yet. But it was there, listed.

I tweaked the settings for the drives in every imaginable way. Eventually, I had hat IDE 0 Master (the first drive listed) set for Auto and LBA. I had IDE0 Slave set to Auto/Auto and IDE1 Master set to Auto/Auto.

I think I had the controller set to IDE mode.

It booted. I was able to format the new drive from Windows. I did a reboot from Windows and it has not booted since.

I am totally lost and incredibly frustrated. I did this many times in the past with real IDE drives with no problem. Of course, back then you had to set a jumper on the drive to set master/slave but my understanding of SATA is that there really is no master slave any more.

Now, I can't get it to boot with just the old drive installed.

Help?

It goes through POST OK and eventually stops with "DISK BOOT FAILURE, INSERT SYSTEM DISK AND PRESS ENTER". I suppose it's possible that the disk really won't boot but it did boot once. I've tried all kinds of settings in the BIOS and nothing seems to work.

The CD shows on on IDE Channel 1 Master, set to Auto/LBA. It is plugged into the second SATA slot, SATA2-1. The HD is plugged into SATA2-0. I also set it for Auto/LBA. During post, I get a message "Detecting IDE drives . . ."
That stays on screen for a long long time, minutes.

I press escape and get a "Prepare to Enter Setup . . ." and it stays on screen for minutes as well.

I am so lost.

I do have a very recent backup of all documents and if I boot into Paragon with the new drive plugged in, it can see it and could restore it to the new drive but it's not the OS and all the software.

I have an image but it's on the nonworking drive. Doh. Creating a new image on the new drive was really my goal today but I only had that one shining moment when both drives were functioning and I formatted the new one. Now I can't seem to access the old one no matter what.

The new drive works so easily that I am becoming convinced that the old drive really is failing and may not ever work again. I'll let it cool off and try again later. That is one thing that had happened when it worked earlier. The system had been powered off completely for about 90 minutes. If I can get the old drive to come back for just a little while, I can copy the system image to the new drive or to an external drive and salvage it. I also have a Win7 recovery disk but it's pretty out of date. Not years or anything like that but a few months.

Anybody got any good ideas. What are the correct settings for the hard disk system in the Bios. What I have now is working well enough to boot from the CD and the restore utility can see the new drive.

The thing is, going back to the beginning, the computer was working fine. I powered it down, unplugged it, installed the new drive, plugged it into SATA2-2 (old drive was in 2-0, CD in 2-1) and turned it on. It would never boot in that set up. The time I got it to boot and formatted the new drive, I think the CD was unplugged and the two HDs were in -0 and -1. That doesn't work any more.

[wallypiper]

Re: Some problems, some symptoms, one clue

Wow.
I decided, before going to bed, to try plugging the old drive into a different SATA socket. It's always been in SATA2-0.
I put it in SATA2-5 and booted from the Paragon CD.
The boot was quick, Paragon immediately saw both hard drives and is now making an image of the primary partition from old drive on the new drive. As soon as it finishes, I'll see if the machine will boot from the old drive instead of the CD.