Re: wal_buffers, redux

On Sun, Mar 11, 2012 at 12:55 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> I've finally been able to run some more tests of the effect of
> adjusting wal_buffers to values higher than 16MB.  I ran the test on
> the 16 core (x 4 hw threads/core) IBM POWER7 machine, with my usual
> configuration settings:
>
> shared_buffers = 8GB
> maintenance_work_mem = 1GB
> synchronous_commit = off
> checkpoint_segments = 300
> checkpoint_timeout = 15min
> checkpoint_completion_target = 0.9
> wal_writer_delay = 20ms
>
> I ran three 30-minute tests at scale factor 300 with wal_buffers set
> at various values from 16MB up to 160MB, in multiples of 16MB, using
> pgbench with 32 clients and 32 threads in each case.  The short
> version is that 32MB seems to be significantly better than 16MB, by
> about 1000 tps, and after that it gets murky; full results are below.

Currently the max of wal_buffers is 16MB (i.e., the size of one WAL file)
when it's set to -1. Thanks to your result, we should increase the max to
32MB?

Regards,

--
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center