icelava.net

why be normal?
Welcome to icelava.net Sign in | Help
in Search

Machine keeps hanging on disk activity

Last post 08-04-2004, 17:15 by icelava. 6 replies.
Sort Posts: Previous Next
  •  07-30-2004, 11:35 478

    Machine keeps hanging on disk activity

    Ok i have a problem that has been lingering for weeks and still can't isolate the exact component and determine the fault. My server (machine #3) has been hanging, the phenomenon is such that the video shut off, and the machine will freeze up so badly even the reset button won't work - it has to be powered down first.

    What I do know is it's always related to an intensive amount of disk activity. E.g. installing software, BitTorrent file checks, etc. As of late, it's been happening more regularly. Surprisingly, huge file transfers between drives and network hadn't caused any freeze.

    This happens regardless of whether hard disks are attached to the onboard IDE (ATA66) controller, the Medley ATA133 controller, or QLogic SCSI controller. I also swapped many IDE cables all around, and even plugged out those expansion controllers to no avail, or even without the optical drives.

    I have even cloned the original hard disk (which had bad sectors) to a newer bigger model. I have also flashed the BIOS of the motherboard. It was originally 256MB of RAM, but I've played with either slab of RAM now and it ain't helpful.

    The OS is Windows Server 2003 EE Eval. I could install that the first time on the ATA133 controller with no problems. But as of now I've gone through over 10 times of failed repair installations with different hardware combos. Now I'm stuck with an unoperational system.
  •  07-30-2004, 13:18 479 in reply to 478

    Re: Machine keeps hanging on disk activity

    Additional note: i was advised to use memtest86 to test the RAM. It's taking very long to run the tests (so far no errors). The odd thing is the PC133 RAM throughout is rated at 310MB/s - isn't the theoretical rating 1064MB/s? Kinda slow don't you think? This is rather odd as system operations in Windows 2003 were rather speedy.

    Update: memtest86 has reported on Test 2, Pass 2, a failure.
    Failing addr: 00000000410 - 0.0MB
    Good: 00000410
    Bad: 00000414
    Err-Bits: 00000004
    Count: 1

    Still trying to make sense of that....

    Update: That could've been a fluke. Nothing happened on the first and third passes. I've been advised to use Windows Memory Diagnostic while at it - a similar memory checker tool. I ran the standard suite of tests on one pass and it appears fine. Running the Extended test suite now.

    Update: Microsoft's WMD appears to come out very clean - no errors. I don't suspect it's a RAM-only problem - checking memory alone doesn't involve the disk activity necessary to induce that problem. Immediately after I quit the check, i tried installing Windows again and it bombed. It must be a component (on the motherboard) that deals with disk IO regardless of controller. I am running memtest86 again but I believe that'll be in vain.
  •  07-30-2004, 23:33 481 in reply to 478

    Re: Machine keeps hanging on disk activity

    Extra things (don't say I didn't try):

    1. swap graphic card - no joy.

    2. plug in another hard disk to install Windows XP - bombs in text-based setup while copying files.

    3. removed CPU. washed off existing thermal tape, and applied new layer of thermal paste. Obviously it ain't a heat problem since it remains at 55ºC and still crashes.

    I am figuring what in the motherboard operates whenever there is disk IO, regardless of onboard or plug-in controllers (IDE/SCSI).

    Yes, I've spent the whole of last night through to dawn and clocked 4 hours of sleep before wasting the whole of today trying out different combos. I'm about to buy a new machine....
  •  07-31-2004, 0:11 482 in reply to 478

    CAUSE DISCOVERED?

    Ok, i was feeling hot, and decided to switch on the air conditioner before bathing. I come back and try again. This time the repair installation went through the final stages and rebooted back to the original Windows state I previously had.

    This is what I infer: it's (certainly) not the CPU that was getting too hot but rather, the motherboard chipset. That kinda makes sense since the chipset has to work hard to direct traffic from disk IO. A mystery is why file transfers don't cause any problem.

    The casing already has a set of inlet/outlet fan, but has a spare outlet slot. Looks like the next step I can take is to install another fan on that spare slot.
  •  07-31-2004, 12:21 483 in reply to 482

    I AM SO WRONG: cause NOT discovered

    Last nigth's repair installation succeeded as a fluke. The OS was still messed up so I had to install a totally new copy instead. And at the "Saving settings..." phase, it bombed again. And again, and again.

    I got fed up and went to sleep, woke up just now in the "morning mist" of the air conditioner and attempted again. It still bombs. wunderbar. Now I am trying it out a down-version Windows 2000 Server. My expectations (since Windows XP smacked too) isn't gonna be too high.

    Update: Win 2000 text-based setup can't even get pass the "Starting up Windows 2000" message LOL. Nevermind I that, I did manage to install Win 2003 again by plugging in the AthlonXP 1600+ (which down-clocked to 1050Mhz thanks to 100Mhz bus) from my bro's PC. On top of that I had Convergent, the distributor for the Kingston memory, took one of my 256MB modules (CL2 single-sided DIMM) after the service technician indeed found bit errors similar to the memtest86 result.

    This reminds me of the old problem I had with my Abit KG7 + AXP1600+ - it was very unstable and hung frequently. However putting the AXP1600+ into an ASUS board was ok, and plugging in an AXP2000+ with the Abit board was fine. Just some serious personality conflicts between the components, I guess.

    Which brings to wonder if it's the CPU vs motherboard, or CPU vs RAM? I'm using the Thunderbird Athlon again now, and will try some disk activity later on after reformatting my Maxtor 80GB, whose partitions were totally toasted by the incessant installations and stalls.

    Update: Still hangs. :'( while trying to run file check in BitTorrent. Should be the CPU vs motherboard.
  •  08-03-2004, 3:02 494 in reply to 483

    Stability attained

    Many thanks to Jacob Soo for lending his 1Ghz Thunderbird Athlon for me to test further. Guess the verdict's out - my own CPU and mobo are just in bitter "disagreement". With his CPU disk activity is all mighty fine, even with the ever-worrying BitTorrent.

    The problem is, this CPU has a multiplier of 7.5x, meant to work with a 133Mhz FSB. This motherboard, the MSI K7T Pro, has a lousy limit of 100Mhz so it runs 750Mhz :-(
  •  08-04-2004, 17:15 495 in reply to 494

    Re: Stability attained

    Bought another 1.2G Thunderbird and it works great with the motherboard - no hang ups after tests with several huge BitTorrent streams checking simultaneously.

    Now I just need to find a buyer for the old CPU....
View as RSS news feed in XML
Powered by Community Server, by Telligent Systems