Hi Michael, > -(snip)- > The other two ones show all problems. I first thought they were broken disks > (which is still a possibility). I checked them with badblocks. sdb gets stuck > at some point and stops responding. sdc gets through but takes ages. Is this > normal - they are 400G disks. I have turned my attention to sdc, after sdb > segfaulted a cp with huge amounts of data. I noticed that even accesing the > better one, sdc, causes the iowait to go up to 100% for longer periods of > time. > This is on a stock debian kernel 2.6.11-1-686 (apparently I should be using a > smp kernel for hyperthreading support - would this make any difference?). > > Has anyone seem anything similiar, hints? I'm running out of ideas but find it > strange that both disks should be broken. with SATA i've so far had success using noapic and poking libata around. You might want to try using the native SATA and libata SATA support together with noapic on a recompiled kernel. If you're lucky either of these 2 combinations will change your situation. Good luck ;) Hth. -- Best regards, Kilian
Attachment:
signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil