The DM multipath service-time path-selector has historically tracked the
amount of outstanding IO per path and used that to approximate the
service-time of each path. In practice this has shown itself to work
fairly well but we can do better by measuring the actual service-time
during IO completio
On Fri, Apr 08 2016 at 2:58pm -0400,
Mike Snitzer wrote:
> The DM multipath service-time path-selector has historically tracked the
> amount of outstanding IO per path and used that to approximate the
> service-time of each path. In practice this has shown itself to work
> fairly well but we ca
On Fri, Apr 08 2016 at 3:03pm -0400,
Mike Snitzer wrote:
> On Fri, Apr 08 2016 at 2:58pm -0400,
> Mike Snitzer wrote:
>
> > The DM multipath service-time path-selector has historically tracked the
> > amount of outstanding IO per path and used that to approximate the
> > service-time of each