I agree with Hannes on this. The default values are not meant to provide optimal performance ... they are meant to provide an accessible LUN whatever the hardware and on the broader range of Linux kernels. Vendors are responsible to provide their array tuning as hwtable entries. And the size or factorization of the hwtable is not an issue.
On Thu, Aug 11, 2016 at 7:58 AM, Hannes Reinecke <[email protected]> wrote: > On 08/10/2016 06:20 PM, Xose Vazquez Perez wrote: > > service-time is more advanced, and it works better in asymmetric and > > asymmetric environments. > > > > Cc: Hannes Reinecke <[email protected]> > > Cc: Benjamin Marzinski <[email protected]> > > Cc: Christophe Varoqui <[email protected]> > > Cc: device-mapper development <[email protected]> > > Signed-off-by: Xose Vazquez Perez <[email protected]> > > --- > > libmultipath/hwtable.c | 2 -- > > 1 file changed, 2 deletions(-) > > > > diff --git a/libmultipath/hwtable.c b/libmultipath/hwtable.c > > index bc2d407..edfd8fe 100644 > > --- a/libmultipath/hwtable.c > > +++ b/libmultipath/hwtable.c > > @@ -198,7 +198,6 @@ static struct hwentry default_hw[] = { > > { > > .vendor = "XtremIO", > > .product = "XtremApp", > > - .selector = "queue-length 0", > > .pgpolicy = MULTIBUS, > > .pgfailback = FAILBACK_UNDEF, > > }, > > @@ -823,7 +822,6 @@ static struct hwentry default_hw[] = { > > { > > .vendor = "PURE", > > .product = "FlashArray", > > - .selector = "queue-length 0", > > .pgpolicy = MULTIBUS, > > .pgfailback = FAILBACK_UNDEF, > > .fast_io_fail = 10, > > > I would _not_ do this, unless tested on the real hardware. > Just claiming 'the other one is more advanced' is not enough here. > > NACK. > > Cheers, > > Hannes > -- > Dr. Hannes Reinecke Teamlead Storage & Networking > [email protected] +49 911 74053 688 > SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg > GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton > HRB 21284 (AG Nürnberg) >
-- dm-devel mailing list [email protected] https://www.redhat.com/mailman/listinfo/dm-devel
