From 1021bd6a89bc756c2a72b29eeee1a7a5c25f64f5 Mon Sep 17 00:00:00 2001
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
Date: Fri, 3 Oct 2014 13:01:27 -0300
Subject: [PATCH] Don't balance vacuum cost delay when per-table settings are
 in effect

When there are cost-delay-related storage options set for a table,
trying to make that table participate in the autovacuum cost-limit
balancing algorithm produces undesirable results: instead of using the
configured values, the global values are always used,
as illustrated by Mark Kirkwood in
http://www.postgresql.org/message-id/52FACF15.8020507@catalyst.net.nz

Since the mechanism is already complicated, just disable it for those
cases rather than trying to make it cope.  There are undesirable
side-effects from this too, namely that the total I/O impact on the
system will be higher whenever such tables are vacuumed.  However, this
is seen as less harmful than slowing down vacuum, because that would
cause bloat to accumulate.  Anyway, in the new system it is possible to
tweak options to get the precise behavior one wants, whereas with the
previous system one was simply hosed.

This has been broken forever, so backpatch to all supported branches.
This might affect systems where cost_limit and cost_delay have been set
for individual tables.
---
 doc/src/sgml/maintenance.sgml       |  9 ++++++---
 src/backend/postmaster/autovacuum.c | 23 ++++++++++++++++++++---
 2 files changed, 26 insertions(+), 6 deletions(-)

diff --git a/doc/src/sgml/maintenance.sgml b/doc/src/sgml/maintenance.sgml
index d692308cf49..8764e0091ac 100644
--- a/doc/src/sgml/maintenance.sgml
+++ b/doc/src/sgml/maintenance.sgml
@@ -792,10 +792,13 @@ analyze threshold = analyze base threshold + analyze scale factor * number of tu
    </para>
 
    <para>
-    When multiple workers are running, the cost limit is
+    When multiple workers are running, the cost delay parameters are
     <quote>balanced</quote> among all the running workers, so that the
-    total impact on the system is the same, regardless of the number
-    of workers actually running.
+    total I/O impact on the system is the same regardless of the number
+    of workers actually running.  However, any workers processing tables whose
+    <literal>autovacuum_vacuum_cost_delay</> or
+    <literal>autovacuum_vacuum_cost_limit</> have been set are not considered
+    in the balancing algorithm.
    </para>
   </sect2>
  </sect1>
diff --git a/src/backend/postmaster/autovacuum.c b/src/backend/postmaster/autovacuum.c
index c240d2444c8..1d6e3f35f94 100644
--- a/src/backend/postmaster/autovacuum.c
+++ b/src/backend/postmaster/autovacuum.c
@@ -193,6 +193,7 @@ typedef struct autovac_table
 	int			at_multixact_freeze_table_age;
 	int			at_vacuum_cost_delay;
 	int			at_vacuum_cost_limit;
+	bool		at_dobalance;
 	bool		at_wraparound;
 	char	   *at_relname;
 	char	   *at_nspname;
@@ -223,6 +224,7 @@ typedef struct WorkerInfoData
 	Oid			wi_tableoid;
 	PGPROC	   *wi_proc;
 	TimestampTz wi_launchtime;
+	bool		wi_dobalance;
 	int			wi_cost_delay;
 	int			wi_cost_limit;
 	int			wi_cost_limit_base;
@@ -1716,6 +1718,7 @@ FreeWorkerInfo(int code, Datum arg)
 		MyWorkerInfo->wi_tableoid = InvalidOid;
 		MyWorkerInfo->wi_proc = NULL;
 		MyWorkerInfo->wi_launchtime = 0;
+		MyWorkerInfo->wi_dobalance = false;
 		MyWorkerInfo->wi_cost_delay = 0;
 		MyWorkerInfo->wi_cost_limit = 0;
 		MyWorkerInfo->wi_cost_limit_base = 0;
@@ -1776,17 +1779,19 @@ autovac_balance_cost(void)
 	if (vac_cost_limit <= 0 || vac_cost_delay <= 0)
 		return;
 
-	/* caculate the total base cost limit of active workers */
+	/* calculate the total base cost limit of participating active workers */
 	cost_total = 0.0;
 	dlist_foreach(iter, &AutoVacuumShmem->av_runningWorkers)
 	{
 		WorkerInfo	worker = dlist_container(WorkerInfoData, wi_links, iter.cur);
 
 		if (worker->wi_proc != NULL &&
+			worker->wi_dobalance &&
 			worker->wi_cost_limit_base > 0 && worker->wi_cost_delay > 0)
 			cost_total +=
 				(double) worker->wi_cost_limit_base / worker->wi_cost_delay;
 	}
+
 	/* there are no cost limits -- nothing to do */
 	if (cost_total <= 0)
 		return;
@@ -1801,6 +1806,7 @@ autovac_balance_cost(void)
 		WorkerInfo	worker = dlist_container(WorkerInfoData, wi_links, iter.cur);
 
 		if (worker->wi_proc != NULL &&
+			worker->wi_dobalance &&
 			worker->wi_cost_limit_base > 0 && worker->wi_cost_delay > 0)
 		{
 			int			limit = (int)
@@ -1815,12 +1821,14 @@ autovac_balance_cost(void)
 			worker->wi_cost_limit = Max(Min(limit,
 											worker->wi_cost_limit_base),
 										1);
+		}
 
-			elog(DEBUG2, "autovac_balance_cost(pid=%u db=%u, rel=%u, cost_limit=%d, cost_limit_base=%d, cost_delay=%d)",
+		if (worker->wi_proc != NULL)
+			elog(DEBUG2, "autovac_balance_cost(pid=%u db=%u, rel=%u, dobalance=%s cost_limit=%d, cost_limit_base=%d, cost_delay=%d)",
 				 worker->wi_proc->pid, worker->wi_dboid, worker->wi_tableoid,
+				 worker->wi_dobalance ? "yes" : "no",
 				 worker->wi_cost_limit, worker->wi_cost_limit_base,
 				 worker->wi_cost_delay);
-		}
 	}
 }
 
@@ -2284,6 +2292,7 @@ do_autovacuum(void)
 		LWLockAcquire(AutovacuumLock, LW_EXCLUSIVE);
 
 		/* advertise my cost delay parameters for the balancing algorithm */
+		MyWorkerInfo->wi_dobalance = tab->at_dobalance;
 		MyWorkerInfo->wi_cost_delay = tab->at_vacuum_cost_delay;
 		MyWorkerInfo->wi_cost_limit = tab->at_vacuum_cost_limit;
 		MyWorkerInfo->wi_cost_limit_base = tab->at_vacuum_cost_limit;
@@ -2579,6 +2588,14 @@ table_recheck_autovac(Oid relid, HTAB *table_toast_map,
 		tab->at_relname = NULL;
 		tab->at_nspname = NULL;
 		tab->at_datname = NULL;
+
+		/*
+		 * If any of the cost delay parameters has been set individually for
+		 * this table, disable the balancing algorithm.
+		 */
+		tab->at_dobalance =
+			!(avopts && (avopts->vacuum_cost_limit > 0 ||
+						 avopts->vacuum_cost_delay > 0));
 	}
 
 	heap_freetuple(classTup);
-- 
GitLab