Unstable Rig - how to diagnose

[beeph]

4 x 6870's
biostar 890FXA
hextec 1050W power supply
2 risers for the gfx cards

stock voltage/clocks
default fan control on ubuntu 11.04

i got 4 rigs running and this particular one always seems to die on me, hard crash 4-48 hours after rebooting.  Problem is it runs for at least a few hours so im never able to catch it int the act. 

Anyone know how to diagnose this problem?  I'm looking for tips/tricks something beyond standard trial and error.. maybe some kind of monitor program, crash log file i dont know about, whatever.

[1715205989]

1715205989

[1715205989]

1715205989

[gyzer]

Are you using 64 bit?

[MiningBuddy]

I would say its an overclocking stability issue, maybe one of those cards is of a slightly different version and cant achieve the same high clock.

[zerokwel]

I would go with PSU issue myself. Cards are running stock and its a random crash. If the temps are ok I would normally go with that

[gnaget]

I can tell you I have a couple of rigs that are particularity finicky.  It was a cooling issue for me, I got to the point where I am running an a/c unit right into them, and they don't lock up anymore

[zerokwel]

aye he's right heat can be a cause. with random timed crashes depending how hot your gpu ovens get. airflow on that system could be a little different and its causing the crash.

My mining pc is no longer in the main house. Its amazing how much cooler the house is now.