unknown hardware problem

[RabbieBurns]

Im out of ideas now and need some suggestions.

Got a custom built machine that is having some problems of late.

Its got 8 HDs, 4 on the motherboard channels and 4 on a controller card.

All HDS pass all the manufacturers advanced diagnostic tests.

Memtest passes all tests 100%

The problem is, the system will make a CLUNK sound, and then just die. It stays powered on, but everything is dead (video, network etc)

Its definately not OS related as it happens when booted into various OS and even with OS HD unplugged and booted to LiveCD.

Im thinking of ordering a new PSU (current one is 400w not sure if thats too little)

Gonna try a different gfx card but I dont think that is the problem.

I dont think its overheating either.

Im running out of ideas. Any suggestions welcome please.

[SYNACK]

I would test it with a more powerful PSU, it takes a bit of power to ramp hard drives up to speed and eight is quite a few, the clunk is probably the disks doing an emergency park as they run out of power to spin up fully.

[RabbieBurns]

THe thing is, itll run for X amount of time before the Clunk, anywhere between 15 mins and 2 days.

Ive just ordered a 600W psu so will test that out when it arrives.

[Curt_LFC]

400w power supply for 8 disks seems pretty low to me, would definitely try the 600w and see if that fixes the problem

[RabbieBurns]

ok thanks, hope thats all it is.

I was convinced there was a dodgy disk in there causing it to trip but spent the whole day deep scanning them all and they all passed the tests.

[DarkLight]

If the computer can be used without all the disks in, remove one and test it again to see if it has the same issue. If it does, continue through with the removal of the next disk and place the previous one back in. I've had a similar issue to this with all the disks, even though they passed all tests, the primary disk decided it would rather make a noise and stop functioning. As soon as I disabled it and removed it from the system the computer was fine again.

Then again, it could also easily be a power issue, or a number of other vital components that aren't easily tested except by replacement.

[Jay]

Discs dont use much power, compared with things like Graphics Cards and CPU's so I'd say 400 watt ought to be fine. Of course thats only if its 400 watt sustained and not one of these 400w peak on a good day in Feb, when the ambient temp is below 10c, and the year is a prime number types.

(I once ran up a Qtek 550w on a test bed and got less than 300w out of it when maxed out.)

From memory each drive ought to use up to 25w at startup and less in operation. I think the newer SAS ones use much less, as do most 2.5inch.

[FN-GM]

How about air flow? it could be that the computer is knocking off because its to hot in the case.

[bizzel]

If it is the PSU then dying at random intervals sounds possible, especially if those random intervals are at busy times when all the disks are being accessed.

[RabbieBurns]

I probably missed an important bit of information: that its been running fine since feb with 7 disks, and started doing this about a month ago. Ive since added a new disk.

Airflow is fine, ive used the rounded IDE cables so theres no ribbon cables, its a big deep rackmount case. Massive fan for the HDs, and an extra fan for the CPU.

Ive got it sitting here with the top off and its def not too hot.

[RabbieBurns]

[RabbieBurns]

Put in the new 600W PSU this afternoon.

Will see how things go.

[RabbieBurns]

robert@blah ~ $ uptime
 00:13:36 up 11:12,  1 user,  load average: 0.00, 0.00, 0.00

so far so good

[jcollings]

I love it - you've invented a kind of hard disk Buckaroo - just keep on adding hard drives until the whole thing kicks back and stops working.

Most excellent

[RabbieBurns]

Thanks for the PSU suggestions. Its solid as a rock now

jcollings: Still got space for another 3 disks so will still be adding to it for a while yet