Re: ATA FLUSHCACHE

On 04/24/07 05:12, Daniel O'Connor wrote:
> On Tuesday 24 April 2007 18:14, Ollivier Robert wrote:
>> The main problem that I've found is that with a single disk
>> configuration, there is a timeout then the system recovers but with a
>> dual (or maybe more) disks involved, multiple/current timeouts make
>> the system panic and that is not nice.
>>
>> I suspect that changing the timeout may just hide the real problem.
> 
> Hmm, it is a non fatal error for gjournal, I'm kind of surprised it's a 
> fatal one for ZFS..
> 
> If the cache flush really failed (eg due to a stuffed disk) then there 
> really is a problem, but IMO it is likely this is going to be caught by 
> a read or a write operation very soon.
> 
> If it is a transient error then panicing seems to be about the worse 
> response :)
> 


Doesn't a failed ATA FLUSHCACHE mean that the device could not complete 
it's writing of cached bits to stable storage within the timeout period? 
  That seems to me that the flushcache should be called more frequently 
then, so less writes have to be written out.


Eric