> I'm stuck with the analysis. The last crash happened last night.
I am sure Leipzig is currently experiencing the same series of hot
days we are. Your failures appear to correlate.
They last time one of my (mostly idle) systems would wegde up during
disk operations only a walk into the datacenter and opening the server
enclosure resolved the issue:
It turned out that one of the two plastic push pins holding down
a heatsink on the RAID controller was broken, and the heatsink was
only dangling around, cooling nothing. (Easy fix by using a cable tie
instead of the broken palstic pin and carefully removing the push pin's
pressure spring from the main board.)
The syslog only indicated timeouts / failures on disk block accesses.
Martin
