Stewart, thanks for the feedback. That this was simple for Drizzle is
encouraging to hear.
Is there a proposal somewhere for eliminating the need for fsyncs on a
slave? As I understand, MWL#164 still requires the fsync.
On Mon, Jan 28, 2013 at 8:38 PM, Vadim Tkachenko <vadim@stripped> wrote:
> I would pretty much like to see benchmarks numbers, especially
> under IO-bound workload before agreeing that it performs well.
> On Mon, Jan 28, 2013 at 5:05 PM, Stewart Smith <stewart@stripped>
>> Kristian Nielsen <knielsen@stripped> writes:
>>> You will be writing your data *6* times: To the real table, when flushing
>>> buffer pool pages. To the double-write buffer while flushing, and to the
>>> log before flushing. Then to the binlog table, and before that to the redo
>>> and the doublewrite buffer. I could never reconcile myself with this.
>> Mind you, it performs really quite well due to the reduced fsync()s and
>> (at least in Drizzle code) was rather trivial to implement. Big bang for
>> buck :)
>> Stewart Smith
> Vadim Tkachenko, CTO, Percona
> Phone +1-925-400-7377, Skype: vadimtk153
> Schedule meeting: http://meetme.so/VadimTkachenko
> Join us in Santa Clara for the annual Percona Live MySQL Conference & Expo 2013!
> Looking for Replication with Data Consistency?
> Try Percona XtraDB Cluster!