ataraid's revenge! (Was: Re: A nasty ataraid experience.)

[Bruce Simpson]

Gavin Atkinson wrote:
> On Thu, 2009-07-23 at 19:57 +0100, Bruce Simpson wrote:
>   
>> 6 months on, ataraid(4) did it again.
>>     
>
> Which ataraid(4) controller is this?  Seeing a verbose dmesg would help.
> There are several patches in the PR database for ataraid problems, it
> would be worth having a look.
>   

Not much to report; it is a JMicron PCI-e card. I switched to the 
JMicron because it actually worked. The on-board controller is VIA and I 
don't use it; I've had problems managing the VIA based software RAID.

Occasionally the mirror degrades, usually on boot, if something panics 
the machine. This leads to interesting inconsistencies and panics. All I 
was doing at the time was srm'ing a bunch of sensitive files, and 
running some CPU (not disk) intensive regression tests for Boost. I 
found it's difficult to recover from errors. See my post from 6 months 
ago about how 'atacontrol rebuild' often just plain fails.

As per original post, the process(es) just wedged in getblk, and I had 
to take the system down to get some sense out of it as disk access 
ground to a halt.