Hi, On Mon, 2002-12-16 at 17:20, glen wrote: > > memtest86 is a good first diagnostic in this sort of case. > > ^^^ > I agree - the first time I brought back a customers system with these > problems I put it on memtest86 for over 48 hours but no errors were reported > - really frustrating as I was convinced the memory was at fault. I've seen memory problems where memtest86 ran just fine --- sometimes it simply takes a different access pattern to trigger the fault. Even if the memory itself is fine, you can have chipset problems that corrupt data in unpredictable ways (I've got a system myself where if you insert a radeon graphics card into the AGP slot, then the secondary on-board IDE controller starts randomly corrupting writes by forgetting a 4-byte transfer every so often.) These things are unpredictable --- carefully monitoring the systems affected so that you know which boxes are stable and which are not is just as valuable as anything else. memtest86 is a good first step, but it's only a first step. Cheers, Stephen _______________________________________________ Ext3-users@redhat.com https://listman.redhat.com/mailman/listinfo/ext3-users
