Re: Kernel problem or faulty drive?

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

> At various times the drive seems to get into a state whereby it has repeated
> trouble communicating with the kernel.  The drive clicks once, there's some
> seek activity by the drive and the following set of messages (along with
> variations) are written to the log:
> 
>   ata3: exception Emask 0x50 SAct 0x0 SErr 0x4010800 action 0xe frozen
>   ata3: irq_stat 0x00400040, connection status changed
>   ata3: SError: { HostInt PHYRdyChg DevExch }

It reported the drive being hot plugged.

>   ata3: hard resetting link
>   ata3: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
>   ata3.00: configured for UDMA/33

And a slightly odd choice of speed.

> This can occur a dozen or more times before the drive settles down, or it
> may occur only once.  All this is while Linux is not accessing the drive and
> none of its partitions are mounted.

Check the cables their connections and check your power is sufficient, a
drive dropping off the bus is a classic symptom of insufficient power on
the supply line it is using (with some PSUs moving to a different
'strand' of cable from the PSU box may be enough.