database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-12-06 00:02:13 +03:00
Code Activity

HOT updates. When we update a tuple without changing any of its indexed

Browse Source 
columns, and the new version can be stored on the same heap page, we no longer
generate extra index entries for the new version.  Instead, index searches
follow the HOT-chain links to ensure they find the correct tuple version.

In addition, this patch introduces the ability to "prune" dead tuples on a
per-page basis, without having to do a complete VACUUM pass to recover space.
VACUUM is still needed to clean up dead index entries, however.

Pavan Deolasee, with help from a bunch of other people.
This commit is contained in:
Tom Lane
2007-09-20 17:56:33 +00:00
parent bbf4fdc253
commit 282d2a03dd
65 changed files with 3517 additions and 514 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
src/backend/postmaster/pgstat.c
View File

@@ -13,7 +13,7 @@
*
* Copyright (c) 2001-2007, PostgreSQL Global Development Group
*
* $PostgreSQL: pgsql/src/backend/postmaster/pgstat.c,v 1.163 2007/09/11 03:28:05 tgl Exp $
* $PostgreSQL: pgsql/src/backend/postmaster/pgstat.c,v 1.164 2007/09/20 17:56:31 tgl Exp $
* ----------
*/
#include "postgres.h"
@@ -1294,7 +1294,7 @@ pgstat_count_heap_insert(Relation rel)
* pgstat_count_heap_update - count a tuple update
*/
void
pgstat_count_heap_update(Relation rel)
pgstat_count_heap_update(Relation rel, bool hot)
{
PgStat_TableStatus *pgstat_info = rel->pgstat_info;
@@ -1304,6 +1304,9 @@ pgstat_count_heap_update(Relation rel)
/* t_tuples_updated is nontransactional, so just advance it */
pgstat_info->t_counts.t_tuples_updated++;
/* ditto for the hot_update counter */
if (hot)
pgstat_info->t_counts.t_tuples_hot_updated++;
/* We have to log the transactional effect at the proper level */
if (pgstat_info->trans == NULL ||
@@ -1340,6 +1343,23 @@ pgstat_count_heap_delete(Relation rel)
}
}
/*
* pgstat_update_heap_dead_tuples - update dead-tuples count
*
* The semantics of this are that we are reporting the nontransactional
* recovery of "delta" dead tuples; so t_new_dead_tuples decreases
* rather than increasing, and the change goes straight into the per-table
* counter, not into transactional state.
*/
void
pgstat_update_heap_dead_tuples(Relation rel, int delta)
{
PgStat_TableStatus *pgstat_info = rel->pgstat_info;
if (pgstat_collect_tuplelevel && pgstat_info != NULL)
pgstat_info->t_counts.t_new_dead_tuples -= delta;
}
/* ----------
* AtEOXact_PgStat
@@ -2901,6 +2921,7 @@ pgstat_recv_tabstat(PgStat_MsgTabstat *msg, int len)
tabentry->tuples_inserted = tabmsg[i].t_counts.t_tuples_inserted;
tabentry->tuples_updated = tabmsg[i].t_counts.t_tuples_updated;
tabentry->tuples_deleted = tabmsg[i].t_counts.t_tuples_deleted;
tabentry->tuples_hot_updated = tabmsg[i].t_counts.t_tuples_hot_updated;
tabentry->n_live_tuples = tabmsg[i].t_counts.t_new_live_tuples;
tabentry->n_dead_tuples = tabmsg[i].t_counts.t_new_dead_tuples;
tabentry->blocks_fetched = tabmsg[i].t_counts.t_blocks_fetched;
@@ -2923,6 +2944,7 @@ pgstat_recv_tabstat(PgStat_MsgTabstat *msg, int len)
tabentry->tuples_inserted += tabmsg[i].t_counts.t_tuples_inserted;
tabentry->tuples_updated += tabmsg[i].t_counts.t_tuples_updated;
tabentry->tuples_deleted += tabmsg[i].t_counts.t_tuples_deleted;
tabentry->tuples_hot_updated += tabmsg[i].t_counts.t_tuples_hot_updated;
tabentry->n_live_tuples += tabmsg[i].t_counts.t_new_live_tuples;
tabentry->n_dead_tuples += tabmsg[i].t_counts.t_new_dead_tuples;
tabentry->blocks_fetched += tabmsg[i].t_counts.t_blocks_fetched;
@@ -2931,6 +2953,8 @@ pgstat_recv_tabstat(PgStat_MsgTabstat *msg, int len)
/* Clamp n_live_tuples in case of negative new_live_tuples */
tabentry->n_live_tuples = Max(tabentry->n_live_tuples, 0);
/* Likewise for n_dead_tuples */
tabentry->n_dead_tuples = Max(tabentry->n_dead_tuples, 0);
/*
* Add per-table stats to the per-database entry, too.
@@ -3115,6 +3139,7 @@ pgstat_recv_vacuum(PgStat_MsgVacuum *msg, int len)
else
tabentry->vacuum_timestamp = msg->m_vacuumtime;
tabentry->n_live_tuples = msg->m_tuples;
/* Resetting dead_tuples to 0 is an approximation ... */
tabentry->n_dead_tuples = 0;
if (msg->m_analyze)
{