David Greaves <david <at> dgreaves.com> writes: > > Check out the linux-ide archive for my (and others) reports. I just read them and weeped. > > I've had lots of issues like this - spurious and IMHO incorrect error > messages. Only certain types of disk access cause them - xfs_repair and > rsync seem to tickle it. Interesting. rsync for me. > With 2.6.15 I had lots of *very* scary moments with multiple disk > failures on a raid5 during xfs_repair. > I think it's down to the 'basic' error handling in the libata code and > certain disks/controllers being loose with the protocol. They then > identified problems in 'fua' (IIRC) handling which was pulled for 2.6.16. > > 2.6.16 seems to be much better (fewer 'odd' errors reported and md > doesn't mind) "Fewer" - oh oh. Can you quantify this for me? I seem to be getting around 8-9 /10 days = or averaging about 1 per rsync. I should check to see if the distribution is Poisson like distribution :). I have not lost my raid1 yet. lucky me. I am now very worried. I need to do some testing off line with the new kernel before putting out the fix. Is your experience also with an Asus board with a via vt8237 powered sata connector? Would you suggest I try the promise sata connector instead, before I try to move over to the newer kernel? Or both? Could you give me more of your experience? Thanks, Mitchell Thanks - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
