On 2013-06-13 16:09, Roger Oberholtzer wrote:
On Thu, 2013-06-13 at 14:52 +0100, Dave Howorth wrote:
Carlos E. R. wrote:
Right but performance matters for nothing if the results are incorrect!
Please read up about fsync before encouraging people not to use it. There are lots of links in this thread and in the XFS list discussion.
But it seems all those discussions are about what happens if the system fails. For some uses I can see that this is a valid concern. I do not think it is a problem for me. If the system should happen to fail, loosing the last X minutes of data is the smaller hassle. Repositioning yourself on the road to start again and all the attendant details are more the problem.
Right. I did hear about abusing fsync years ago, from kernel devs probably. Directly out of the horse mouth.
I am happy to say that system failure in a vehicle has been virtually non-existent. And it is a demanding environment. Of course, we have spent a bit of time getting the power clean and steady.
That's quite an accomplishment :-) -- Cheers / Saludos, Carlos E. R. (from 12.3 x86_64 "Dartmouth" at Telcontar)