pgsql: Add an EXPLAIN (BUFFERS) option to show buffer-usage statistics.

Log Message:
-----------
Add an EXPLAIN (BUFFERS) option to show buffer-usage statistics.

This patch also removes buffer-usage statistics from the track_counts
output, since this (or the global server statistics) is deemed to be a better
interface to this information.

Itagaki Takahiro, reviewed by Euler Taveira de Oliveira.

Modified Files:
--------------
pgsql/contrib/auto_explain:
auto_explain.c (r1.9 -> r1.10)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/contrib/auto_explain/auto_explain.c?r1=1.9&r2=1.10)
pgsql/contrib/pg_stat_statements:
pg_stat_statements.c (r1.7 -> r1.8)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/contrib/pg_stat_statements/pg_stat_statements.c?r1=1.7&r2=1.8)
pgsql/doc/src/sgml:
auto-explain.sgml (r1.5 -> r1.6)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/doc/src/sgml/auto-explain.sgml?r1=1.5&r2=1.6)
pgsql/doc/src/sgml/ref:
explain.sgml (r1.47 -> r1.48)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/doc/src/sgml/ref/explain.sgml?r1=1.47&r2=1.48)
pgsql/src/backend/commands:
copy.c (r1.318 -> r1.319)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/commands/copy.c?r1=1.318&r2=1.319)
explain.c (r1.195 -> r1.196)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/commands/explain.c?r1=1.195&r2=1.196)
tablecmds.c (r1.309 -> r1.310)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/commands/tablecmds.c?r1=1.309&r2=1.310)
pgsql/src/backend/executor:
execMain.c (r1.337 -> r1.338)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/executor/execMain.c?r1=1.337&r2=1.338)
execProcnode.c (r1.68 -> r1.69)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/executor/execProcnode.c?r1=1.68&r2=1.69)
functions.c (r1.137 -> r1.138)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/executor/functions.c?r1=1.137&r2=1.138)
instrument.c (r1.22 -> r1.23)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/executor/instrument.c?r1=1.22&r2=1.23)
spi.c (r1.211 -> r1.212)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/executor/spi.c?r1=1.211&r2=1.212)
pgsql/src/backend/storage/buffer:
buf_init.c (r1.82 -> r1.83)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/storage/buffer/buf_init.c?r1=1.82&r2=1.83)
bufmgr.c (r1.252 -> r1.253)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/storage/buffer/bufmgr.c?r1=1.252&r2=1.253)
localbuf.c (r1.87 -> r1.88)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/storage/buffer/localbuf.c?r1=1.87&r2=1.88)
pgsql/src/backend/storage/file:
buffile.c (r1.34 -> r1.35)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/storage/file/buffile.c?r1=1.34&r2=1.35)
pgsql/src/backend/tcop:
postgres.c (r1.575 -> r1.576)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/tcop/postgres.c?r1=1.575&r2=1.576)
pquery.c (r1.132 -> r1.133)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/tcop/pquery.c?r1=1.132&r2=1.133)
pgsql/src/include/commands:
explain.h (r1.43 -> r1.44)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/include/commands/explain.h?r1=1.43&r2=1.44)
pgsql/src/include/executor:
execdesc.h (r1.40 -> r1.41)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/include/executor/execdesc.h?r1=1.40&r2=1.41)
executor.h (r1.164 -> r1.165)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/include/executor/executor.h?r1=1.164&r2=1.165)
instrument.h (r1.20 -> r1.21)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/include/executor/instrument.h?r1=1.20&r2=1.21)
pgsql/src/include/nodes:
execnodes.h (r1.213 -> r1.214)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/include/nodes/execnodes.h?r1=1.213&r2=1.214)
pgsql/src/include/storage:
buf_internals.h (r1.102 -> r1.103)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/include/storage/buf_internals.h?r1=1.102&r2=1.103)
bufmgr.h (r1.121 -> r1.122)
(http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/include/storage/bufmgr.h?r1=1.121&r2=1.122)

On Tue, 2009-12-15 at 04:57 +0000, Robert Haas wrote:
> This patch also removes buffer-usage statistics from the track_counts
> output, since this (or the global server statistics) is deemed to be a
> better interface to this information.

I hope anything being removed is documented, though I don't see that.

Do you mean pg_statio_* tables no longer track heap_blocks* ? That is my
understanding of what track_counts gives us. I think you mean
log_*_stats, which is something else entirely.

--
Simon Riggs www.2ndQuadrant.com

Simon Riggs escreveu:
> On Tue, 2009-12-15 at 04:57 +0000, Robert Haas wrote:
>> This patch also removes buffer-usage statistics from the track_counts
>> output, since this (or the global server statistics) is deemed to be a
>> better interface to this information.
>
> I hope anything being removed is documented, though I don't see that.
>
That functionality isn't documented AFAIR. That's why Tom suggested to rip it out.

> Do you mean pg_statio_* tables no longer track heap_blocks* ? That is my
> understanding of what track_counts gives us. I think you mean
> log_*_stats, which is something else entirely.
>
That's it!

--
Euler Taveira de Oliveira
http://www.timbira.com/

On Tue, Dec 15, 2009 at 1:32 PM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
> On Tue, 2009-12-15 at 04:57 +0000, Robert Haas wrote:
>> This patch also removes buffer-usage statistics from the track_counts
>> output, since this (or the global server statistics) is deemed to be a
>> better interface to this information.
>
> I hope anything being removed is documented, though I don't see that.
> Do you mean pg_statio_* tables no longer track heap_blocks* ? That is my
> understanding of what track_counts gives us. I think you mean
> log_*_stats, which is something else entirely.

Oops. You're right, it's log_*_stats, rather than track_counts().
Unfortunately, the format of the log_*_stats output is not documented:
we can't remove this from the list of things that log_*_stats is
documented to print, because there is no such list. That might be
worth correcting, but the lack of pre-existing documentation didn't
seem like something this patch should try to fix.

...Robert