Date: Tue, 7 Mar 2023 12:55:08 -0800
From: Nathan Bossart <nathandbossart@gmail.com>
To: Melanie Plageman <melanieplageman@gmail.com>
Cc: Michael Paquier <michael@paquier.xyz>, pgsql-hackers@postgresql.org
Subject: Re: add PROCESS_MAIN to VACUUM
Message-ID: <20230307205508.GA3314274@nathanxps13>
References: <20230306202734.GA3033389@nathanxps13>
 <CAAKRu_ZJMFrO_tspzoX43h9NyEFS1nS438MoS3Yv7JxK3=jFPw@mail.gmail.com>
 <20230306211337.GA3076909@nathanxps13>
 <20230306220958.rk4b4msmybi6hm56@liskov>
 <20230306224310.GA3165453@nathanxps13>
 <CAAKRu_ZEKEsXDfFbOHRMhEROK5h1KcTcwUhERv5ZE0RQjX82_A@mail.gmail.com>
 <ZAaDPG9KJXqFU2Sr@paquier.xyz>
 <20230307005949.GA3266981@nathanxps13>
 <ZAazRGzyCYoRzy1M@paquier.xyz>
 <CAAKRu_amaA3gDUwsKtipWY_ShGSy_O3D8hwN8aT98LcBVK+yHQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="PEIAKu/WMn1b1Hv9"
Content-Disposition: inline
In-Reply-To: 
 <CAAKRu_amaA3gDUwsKtipWY_ShGSy_O3D8hwN8aT98LcBVK+yHQ@mail.gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/20230307205508.GA3314274%40nathanxps13>
Precedence: bulk


--PEIAKu/WMn1b1Hv9
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Tue, Mar 07, 2023 at 12:39:29PM -0500, Melanie Plageman wrote:
> Yes, sounds clear to me also!

Here is an updated patch.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com

--PEIAKu/WMn1b1Hv9
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment;
	filename="process_main_readability_fix_v3.patch"

diff --git a/src/backend/commands/vacuum.c b/src/backend/commands/vacuum.c
index 580f966499..0acc42af2b 100644
--- a/src/backend/commands/vacuum.c
+++ b/src/backend/commands/vacuum.c
@@ -2058,25 +2058,34 @@ vacuum_rel(Oid relid, RangeVar *relation, VacuumParams *params, bool skip_privs)
 	save_nestlevel = NewGUCNestLevel();
 
 	/*
-	 * Do the actual work --- either FULL or "lazy" vacuum
+	 * If PROCESS_MAIN is set (the default), it's time to vacuum the main
+	 * relation.  Otherwise, we can skip this part.  If processing the TOAST
+	 * table is required (e.g., PROCESS_TOAST is set), we'll force PROCESS_MAIN
+	 * to be set when we recurse to the TOAST table so that it gets processed
+	 * here.
 	 */
-	if ((params->options & VACOPT_FULL) &&
-		(params->options & VACOPT_PROCESS_MAIN))
+	if (params->options & VACOPT_PROCESS_MAIN)
 	{
-		ClusterParams cluster_params = {0};
+		/*
+		 * Do the actual work --- either FULL or "lazy" vacuum
+		 */
+		if (params->options & VACOPT_FULL)
+		{
+			ClusterParams cluster_params = {0};
 
-		/* close relation before vacuuming, but hold lock until commit */
-		relation_close(rel, NoLock);
-		rel = NULL;
+			/* close relation before vacuuming, but hold lock until commit */
+			relation_close(rel, NoLock);
+			rel = NULL;
 
-		if ((params->options & VACOPT_VERBOSE) != 0)
-			cluster_params.options |= CLUOPT_VERBOSE;
+			if ((params->options & VACOPT_VERBOSE) != 0)
+				cluster_params.options |= CLUOPT_VERBOSE;
 
-		/* VACUUM FULL is now a variant of CLUSTER; see cluster.c */
-		cluster_rel(relid, InvalidOid, &cluster_params);
+			/* VACUUM FULL is now a variant of CLUSTER; see cluster.c */
+			cluster_rel(relid, InvalidOid, &cluster_params);
+		}
+		else
+			table_relation_vacuum(rel, params, vac_strategy);
 	}
-	else if (params->options & VACOPT_PROCESS_MAIN)
-		table_relation_vacuum(rel, params, vac_strategy);
 
 	/* Roll back any GUC changes executed by index functions */
 	AtEOXact_GUC(false, save_nestlevel);

--PEIAKu/WMn1b1Hv9--