From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Fri, 14 Jun 2019 12:20:01 +0300 From: Vladimir Davydov Subject: Re: [tarantool-patches] [PATCH v3 09/14] wal: a dedicated wal scheduling fiber Message-ID: <20190614092001.krk64zm2ivhudynn@esperanza> References: <64f63cc1cea7a3c632561f36eb043cb386c4e472.1560112747.git.georgy@tarantool.org> <20190613142452.j63yqr4q53xszo4r@esperanza> <3143671.OOzuiaNZFI@home.lan> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <3143671.OOzuiaNZFI@home.lan> To: =?utf-8?B?0JPQtdC+0YDQs9C40Lkg0JrQuNGA0LjRh9C10L3QutC+?= Cc: tarantool-patches@freelists.org List-ID: On Thu, Jun 13, 2019 at 10:36:30PM +0300, Георгий Кириченко wrote: > On Thursday, June 13, 2019 5:24:52 PM MSK Vladimir Davydov wrote: > > On Sun, Jun 09, 2019 at 11:44:38PM +0300, Georgy Kirichenko wrote: > > > > > This adds an extra ctxsw to the relatively hot WAL writer wakeup path so > > we can't commit it. I assume it's a temporary hack you added so that you > > can submit this patches for review. > This adds at most only one ctxsw on each wal branch and we could ask Alexander > to bench it. > > > > I understand why you're doing this, but the right way would be removing > > yields from on_commit triggers. We need to make vy_log_write yield-free. > I'm totally agreed, however, I'm not sure could disable yields for on_commit/ > on_rollback right now. I'll try to do that without introducing a separate vylog thread then so as not to trigger Kostja. If I fail, we will have no choice but fall back on a separate fiber for WAL wakeups.