Re: [techtalk] System CRASH - one solution

2000-01-13 Thread Jeff Dike
> It is always the same inode-- When fscking, inode 818763 is always > deleted with zero dtime. What does this show? This could just be a temporary file that some daemon has opened and unlinked. When the system crashes, the one reference to that file (the process) has disappeared, so fsck se

Re: [techtalk] System CRASH - one solution

2000-01-13 Thread Jennifer Tippens
How about this, though. It is always the same inode-- When fscking, inode 818763 is always deleted with zero dtime. What does this show? I ran badblocks on /dev/hda1 (where the inode would be) and the system once again crashed. Thanks, Jen Robert B Benson wrote: > Greetings all, > > With inter

[techtalk] System CRASH - one solution

2000-01-12 Thread Robert B Benson
Greetings all, With intermittent faults, memory (SIMMs, DIMMS, etc.) is the first place to look. Next, large scale chips like the PCI bus manager (device that manages all) could be the fault. The fault behind the cause is of course HEAT. Heat sinks and more air movement around heat sources coul