On Mon, Aug 6, 2012 at 2:15 PM, Stefan Ring <stefan...@gmail.com> wrote:
> So you're saying that SSDs don't generally flush data to stable medium
> when instructed to? So data written before an fsync is not guaranteed
> to be seen after a power-down?

It depends on the model. Consumer models are less likely to
immediately flush. My understanding that this is done in part to do
some write coalescing and reduce the number of P/E cycles. Enterprise
models should either flush, or contain a super capacitor that provides
enough power for the drive to complete writing any date in its buffer.

> If that -- ignoring cache flush requests -- is the whole reason why
> SSDs are so fast, I'm glad I haven't got one yet.

They're fast for random reads and writes because they don't have seek
latency. They're fast for sequential IO because they aren't limited by
spindle speed.

Brandon High : bh...@freaks.com
zfs-discuss mailing list

Reply via email to