System Locks up and then unlocks from being locked up randomly

[DaveLembke]

My Kabini Sempron 3850 1.3Ghz quadcore workstation is back to acting up again.

HDD LED solid at times and system acting unresponsive ( yet mouse moves freely about 90% of the time when this happens ). And then it will fix itself and start behaving normal again after about a minute or so. Hard Drive is silent not reading any data or writing etc, its a system hang condition until it gets past this hung state on its own after a period of time passes. Other times it locks up tight requiring a power button held in to shut it down.

Memtest86 shows no problems. Windows Event Logs show no problems. CrystalDiskInfo shows no problems. Power supply voltages are correct and not floating around.

This system I shared once before here and it was doing the same thing on a SSD drive, although with the SSD drive the system was far more unstable. I ended up installing Windows 7 64-bit then to a 250GB HDD and the system improved, so i assumed that the SSD I had in it had issues. However now it looks like its something else because the problem is back. However far more infrequent than when I had the SSD in it. When the SSD was in it, it was acting as if the Paging was getting messed up as for getting rid of the virtual memory on the system made it more stable with the SSD. But it made the system slower not being able to page and everything having to juggle in the 4GB DDR3 1600Mhz RAM.

Now on occasion I hear what sounds like a bell ping sound that comes from the hard drive, however Crystaldiskinfo shows that SMART data shows drive is healthy. This ping sound is very infrequent, maybe in a 8 hr period of use I might hear it 1 or 2 times and the noise doesnt happen when the system acts up. I created a share on my other system and so I am mirroring data manually between two systems so that if this 250GB HDD dies I still have most of my data on other system. The ping noise that happens infrequent isnt a noise that I see as healthy.

Anyone seen a Hard Drive die and show signs that its dying and not trip SMART and Windows Event Logs?   Thinking that I might image this drive to a 300GB drive that I have and see if this fixes issue, but its very strange that this system is the second go around with same issue this time HDD vs SSD.

Additionally the lock ups and HDD solid LED happens more when weighing down the system multitasking, yet the same multitasking at times has no problems. To me it points back to HDD issue, but why isnt the system detecting this issue and showing issues when it no longer takes input but mouse pointer can move all over the screen. Problem would be easier to diagnose if I get the same results given the same multitask or work thrown at it.

[BC_Programmer]

S.M.A.R.T can tell you if a drive is failed or failing. It cannot tell you if it's not failed/failing.

I'd suggest you run a chkdsk /r to do a surface scan to verify there are no bad sectors/clusters.

[patio]

You may also wanna swap a different HDD in to test for 72 hrs...if same symptoms it could be the MBoard...

[DaveLembke]

Good suggestion BC on the checkdisk... will give that a shot to see what it reports.

Additionally, i am hoping its not the motherboard, however the motherboard was a very cheap Gigabtye board. The Socket AM1 boards are so cheap. I paid $39.99 with free shipping for this one:  GIGABYTE GA-AM1M-S2H AM1         https://www.newegg.com/Product/Product.aspx?Item=N82E16813128696

It was a very cheap to build computer with CPU costing me $32, motherboard $40, and 4GB RAM for $20. Used an old Pentium 4 case and a newer 350 watt power supply, and other parts on hand that are ( ?? were ?? ) healthy parts.

[BC_Programmer]

I used a similar Gigabyte AM1M-S2P, and it doesn't have problems like that, so it's not inherent to the board. (The only difference between the S2H and the S2P looks like the S2H has HDMI and no Serial/Parallel Ports).

Remember also that  the Motherboard is a "dumb" connection;  it has the Host Adapter but all the work is done by the drive logic since that is where the controller(s) are. So the sort of fialure that would mess up drives would tend to be far more pronounced and less sporadic, I'd suspect.

[Geek-9pm]

A PC motherboard differs a lotfrom those used incomercial servers. Still, the kind of hardware used is very much the xsame stuff.
Of generalinterst is this reportfrom Google.
http://www.datacenterknowledge.com/archives/2008/05/30/failure-rates-in-google-data-centers/
Failure Rates in Google Data Centers

Google’s data centers operate on a large enough scale that it can shift redundancy from hardware to software. There’s so much hardware that it can use software to identify failures and route around them, even at the cluster level.

Yes, hard drives are the worst. But they report someo ther faiures that you would not guess.  Thepoint hee is that some errors are hard to identify . Motherbaord failures aein that class.

[DaveLembke]

- Ran Check Disk and no problems found.

- Downloaded Seatools for Windows since the drive is a Seagate and ran the long test on it and no problems found as well as the system didnt lock up during this period either and acted healthy.

- Downloaded Performance Monitor from Hexagora which has a Hard Drive activity graph to see if there is actually hard drive activity seen by the OS when the HDD LED goes solid and system becomes unresponsive other than mouse movement etc. http://www.hexagora.com/en_dw_davperf.asp

*Going to keep Performance Monitor running and next time it locks up and is unresponsive but mouse movement still able to move, see if the graph shows actual hard drive activity or if nothing is showing on the graph but graph might still be running. While this isnt a solution I am really curious if the hard drive is actually active or if the HDD LED is just showing a solid lit state because of the problem the hardware has. If I see hard drive activity on the graph to support the HDD LED being solid as ok something is going on thats bringing the system to its knees then I can go in that direction, but if the graph shows that its logging but no activity and the HDD LED is solid then I might be dealing with a Motherboard failure, however I would swap out my hard drive first just to be sure that everything points back to the main board as the issue.

Also not discussed prior but to mention now, the CPU/GPU temps are all fine so its not overheating and locking up. Figured I'd update this to let you all know where I am at with this and if there are any other suggestions I will give them a try. Starting to get a sense that well I got what i paid for, cheap computer likely is cheap and wont last, but Gigabyte motherboard while they do have some boards that die, I didnt expect after about a year of somewhat heavy use for it to crap out on me. I like the fact that it doesnt use much electricity and is plenty of processing power to do most of what I do including light gaming.

If all else fails I will decommission this system and put a laptop in place of it that is an electron sipper. I have been eyeballing my one Core i5 laptop with thoughts of placing its guts inside a minitower case. Its the Lenovo that i have the Pentium 4 massive heatsink attached to because it use to have an overheating issue which is why i got it for free, and keyboard set to side so the heatsink is attached to top of the CPU. I can hide the ugly mod in a desktop case and run it as a desktop computer. Solder the soft power button wires to the power push button of laptop and run all the cables through the open card slots at the rear of it. I can use a USB/IDE adapter to use a desktop 5.25" DVD ROM and a power brick that I have that makes a single P-connector to power the Desktop DVD-RW IDE drive. Just havent put the time into it yet to make a new frankenstein box. 

[BC_Programmer]

- Ran Check Disk and no problems found.

Just to confirm, you  did run it with the /r switch, right?

[DaveLembke]

yup, it required a reboot to perform this chkdsk /r at the next boot. It required for it to be a scheduled process for Windows 7 and couldnt be done with 7 running as it was.