The pg_buffercache module provides a means for
examining what's happening in the shared buffer cache in real time.
The module provides a C function pg_buffercache_pages
that returns a set of records, plus a view
pg_buffercache that wraps the function for
convenient use.
By default use is restricted to superusers and members of the
pg_read_all_stats role. Access may be granted to others
using GRANT.
pg_buffercache ViewThe definitions of the columns exposed by the view are shown in Table F.15.
Table F.15. pg_buffercache Columns
| Name | Type | References | Description |
|---|---|---|---|
bufferid | integer | ID, in the range 1..shared_buffers | |
relfilenode | oid | pg_class.relfilenode | Filenode number of the relation |
reltablespace | oid | pg_tablespace.oid | Tablespace OID of the relation |
reldatabase | oid | pg_database.oid | Database OID of the relation |
relforknumber | smallint | Fork number within the relation; see
include/common/relpath.h | |
relblocknumber | bigint | Page number within the relation | |
isdirty | boolean | Is the page dirty? | |
usagecount | smallint | Clock-sweep access count | |
pinning_backends | integer | Number of backends pinning this buffer |
There is one row for each buffer in the shared cache. Unused buffers are
shown with all fields null except bufferid. Shared system
catalogs are shown as belonging to database zero.
Because the cache is shared by all the databases, there will normally be
pages from relations not belonging to the current database. This means
that there may not be matching join rows in pg_class for
some rows, or that there could even be incorrect joins. If you are
trying to join against pg_class, it's a good idea to
restrict the join to rows having reldatabase equal to
the current database's OID or zero.
When the pg_buffercache view is accessed, internal buffer
manager locks are taken for long enough to copy all the buffer state
data that the view will display.
This ensures that the view produces a consistent set of results, while not
blocking normal buffer activity longer than necessary. Nonetheless there
could be some impact on database performance if this view is read often.
regression=# SELECT c.relname, count(*) AS buffers
FROM pg_buffercache b INNER JOIN pg_class c
ON b.relfilenode = pg_relation_filenode(c.oid) AND
b.reldatabase IN (0, (SELECT oid FROM pg_database
WHERE datname = current_database()))
GROUP BY c.relname
ORDER BY 2 DESC
LIMIT 10;
relname | buffers
---------------------------------+---------
tenk2 | 345
tenk1 | 141
pg_proc | 46
pg_class | 45
pg_attribute | 43
pg_class_relname_nsp_index | 30
pg_proc_proname_args_nsp_index | 28
pg_attribute_relid_attnam_index | 26
pg_depend | 22
pg_depend_reference_index | 20
(10 rows)
Mark Kirkwood <markir@paradise.net.nz>
Design suggestions: Neil Conway <neilc@samurai.com>
Debugging advice: Tom Lane <tgl@sss.pgh.pa.us>