Comment 91 for bug 285892

In the end, it seems that this bug is related to
- HDD problems like unreadable blocks / sectors
- possibly changes in how HDDs act in such situations
- some controllers like the one in AMD SB700
- kernel / libata changes after 2.6.24 (?)

I worry that maybe
- HDD manufacturers are changing the way drives act on hardware problems
- relocating sectors takes more time than intended and/or is producing unexpected states in controller and/or kernel driver
- kernel driver does not handle this situations properly

So the appropriate suggestion would be to
- check HDD thoroughly, check SMART state, check seek times, check reading and writing speed / throughput (!)
- replace the HDD with a "RAID-ready" HDD - this devices limit the time for relocating sectors or whatever "self-healing" they do
- having a new HDD, write to all sectors before using it, e.g. dd if=/dev/zero of=/dev/sdf bs=256K
- press kernel / libata developers to investigate the problem

I had the problem with SB700 and SATA WD VelociRaptor 150GB and WD Green 2 TB drives.

best regards + best luck