Re: ad WRITE_DMA timing out frequently

On 2/18/2005, "Colin Percival" <cperciva_at_freebsd.org> wrote:

>Reid Linnemann wrote:
>> smartctl doesn't report any logged errors. On a hunch, I also dd'ed a
>> file large enough to fill /var, hoping that it would crater on writing
>> to that sector. It didn't. I know that's not a very useful test, but
>> it seems to hint to me that the disk isn't bad, but the driver is
>> freaking out from some event.
>
>It's quite possible that the driver is at fault, but I'd run a
>"smartctl -t long" test first, just to make sure the drive isn't
>suffering an intermitant fault.
>
>Colin Percival

I ran smartctl -t long on the disk, and sure enough it's healthy:

SMART Self-test log structure revision number 1
Num  Test_Description    Status                  Remaining 
LifeTime(hours)  LBA_of_first_error
# 1  Extended offline    Completed without error       00%     22444     
   -

There must be some strange system event happening that is causing the
driver to freak out. I really have a hunch it's related to the msp
queue mail that is being dumped in /var/spool/clientmqueue, because the
that occurance roughly falls in lline with the ad0 failures, and it's
all on the same partition that the error is reported on.