From mboxrd@z Thu Jan 1 00:00:00 1970 From: Xose Vazquez Perez Subject: Re: [PATCH 14/15] multipath-tools: replace queue-length by default selector Date: Thu, 11 Aug 2016 18:16:17 +0200 Message-ID: <8aca22a1-73c5-a5e9-e2bd-3180e2eec564@gmail.com> References: <1470846038-6590-1-git-send-email-xose.vazquez@gmail.com> <1470846038-6590-14-git-send-email-xose.vazquez@gmail.com> <411eb480-d21d-051b-1f7b-12631cae1f25@suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <411eb480-d21d-051b-1f7b-12631cae1f25@suse.de> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com To: Hannes Reinecke , Christophe Varoqui , Vincent Chen , Brian Bunker Cc: device-mapper development List-Id: dm-devel.ids On 08/11/2016 07:58 AM, Hannes Reinecke wrote: > On 08/10/2016 06:20 PM, Xose Vazquez Perez wrote: >> service-time is more advanced, and it works better in asymmetric and >> symmetric(corrected) environments. >> - .selector = "queue-length 0", > I would _not_ do this, unless tested on the real hardware. > Just claiming 'the other one is more advanced' is not enough here. service-time works better in any wild environment, different adapters or different links speeds and any kind of load, aka the Production Environment. I guess that manufacturers' tests are done in ideal conditions(PV=nRT). Anyhow, the default path selector algorithm does not depend, by no means, of any array settings or hardware. It's only related to the Linux host and its environment. How PURE/FlashArray and XtremIO/XtremApp were tested? Is there anything wrong with service-time?