From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Wed, 6 Feb 2019 12:20:03 +0300 From: Vladimir Davydov Subject: Re: [tarantool-patches] Re: [PATCH 5/9] vinyl: keep track of dumps per compaction for each LSM tree Message-ID: <20190206092003.mhklof7sqdhbym47@esperanza> References: <32b5bdf6ea829c03d858ca8faffcd11411a113b4.1548017258.git.vdavydov.dev@gmail.com> <20190205165829.GD6811@chai> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20190205165829.GD6811@chai> To: Konstantin Osipov Cc: tarantool-patches@freelists.org List-ID: On Tue, Feb 05, 2019 at 07:58:29PM +0300, Konstantin Osipov wrote: > * Vladimir Davydov [19/01/21 06:58]: > > This patch adds dumps_per_compaction metric to per index statistics. It > > shows the number of dumps it takes to trigger a major compaction of a > > range in a given LSM tree. We need it to automatically choose the > > optimal number of ranges that would smooth out the load generated by > > range compaction. > > I obviously like the idea of dumps_per_compaction :-), but using a > heap to maintain the minimum sounds like a bit of an overkill. Is > average so much less accurate? It would be much cheaper to > maintain. I guess we could do that. I'll try.