On Tue, 2005-03-29 at 08:41 +0100, Rob Kirkbride wrote: > Jess Anderson wrote on 25/03/2005 02:46: > > >Scot L. Harris > > > > > >>Jess Anderson: > >> > >> > >>>I have three slots and three 512MB DIMMs. What I did was run > >>>memtest three times, rotating the DIMMs (call 'em ABC) by one > >>>slot each time: > >>> > >>> > >>Why not pull one DIMM at a time an run memtest86? If you get a > >>clean pass then those two chips are good. > >> > >> > >Ja, that's my plan. > > > > > > > I was getting crashing similar to yourself (often to do with ext3) I had > been grappling with my server starting to crash regularly (dual AMD > machine), ran memtest and everything was reported ok. I decided that my > CPUs were running a little hot so changed the fans on them. This > unfortunately did not cure the problem. I took a stick of ram out and > the server still fell over. I swapped the sticks over and its now been > running for several days (would generally crash once a day before). > Strangely I had a branded stick (crucial) and an unbranded one. Its the > branded one that caused me the problem! > I'd been tearing my hair out on this for ages - very annoying, hope > you're able to solve your problem too! > > Rob > Power supplies that are under-performing or starting to fail can be a real bear as well. I had one that gave me no symptoms except frequent hangs/crashes. I tried memory, cpu, mobo, drives, all to no avail. when I replaced the power supply with one rated exactly the same (350w) the problem magically went away and I went back to my original hardware. That machine is still chuckling happily along. YMMV, but keep checking various things.
