Add temp_file_limit GUC parameter to constrain temporary file space usage.

The limit is enforced against the total amount of temp file space used by
each session.

Mark Kirkwood, reviewed by Cédric Villemain and Tatsuo Ishii

Author: tglsfdc

doc/src/sgml/config.sgml

@@ -1025,10 +1025,41 @@ SET ENABLE_SEQSCAN TO OFF;
      </variablelist>
      </sect2>
 
+     <sect2 id="runtime-config-resource-disk">
+     <title>Disk</title>
+
+     <variablelist>
+     <varlistentry id="guc-temp-file-limit" xreflabel="temp_file_limit">
+      <term><varname>temp_file_limit</varname> (<type>integer</type>)</term>
+      <indexterm>
+       <primary><varname>temp_file_limit</> configuration parameter</primary>
+      </indexterm>
+      <listitem>
+       <para>
+        Specifies the maximum amount of disk space that a session can use
+        for temporary files, such as sort and hash temporary files, or the
+        storage file for a held cursor.
+        The value is specified in kilobytes, and <literal>-1</> (the
+        default) means no limit.
+        Only superusers can change this setting.
+       </para>
+       <para>
+        This setting constrains the total space used at any instant by all
+        temporary files used by a given <productname>PostgreSQL</> session.
+        It should be noted that disk space used for explicit temporary
+        tables, as opposed to temporary files used behind-the-scenes in query
+        execution, does <emphasis>not</emphasis> count against this limit.
+       </para>
+      </listitem>
+     </varlistentry>
+
+     </variablelist>
+     </sect2>
+
      <sect2 id="runtime-config-resource-kernel">
      <title>Kernel Resource Usage</title>
-     <variablelist>
 
+     <variablelist>
      <varlistentry id="guc-max-files-per-process" xreflabel="max_files_per_process">
       <term><varname>max_files_per_process</varname> (<type>integer</type>)</term>
       <indexterm>

src/backend/storage/file/fd.c

@@ -128,9 +128,6 @@ static int	max_safe_fds = 32;	/* default if not changed */
 #define FD_XACT_TRANSIENT	(1 << 2)	/* T = close (not delete) at aoXact,
 										 * but keep VFD */
 
-/* Flag to tell whether there are files to close/delete at end of transaction */
-static bool have_pending_fd_cleanup = false;
-
 typedef struct vfd
 {
 	int			fd;				/* current FD, or VFD_CLOSED if none */
@@ -140,6 +137,7 @@ typedef struct vfd
 	File		lruMoreRecently;	/* doubly linked recency-of-use list */
 	File		lruLessRecently;
 	off_t		seekPos;		/* current logical file position */
+	off_t		fileSize;		/* current size of file (0 if not temporary) */
 	char	   *fileName;		/* name of file, or NULL for unused VFD */
 	/* NB: fileName is malloc'd, and must be free'd when closing the VFD */
 	int			fileFlags;		/* open(2) flags for (re)opening the file */
@@ -159,6 +157,17 @@ static Size SizeVfdCache = 0;
  */
 static int	nfile = 0;
 
+/* True if there are files to close/delete at end of transaction */
+static bool have_pending_fd_cleanup = false;
+
+/*
+ * Tracks the total size of all temporary files.  Note: when temp_file_limit
+ * is being enforced, this cannot overflow since the limit cannot be more
+ * than INT_MAX kilobytes.  When not enforcing, it could theoretically
+ * overflow, but we don't care.
+ */
+static uint64 temporary_files_size = 0;
+
 /*
  * List of stdio FILEs and <dirent.h> DIRs opened with AllocateFile
  * and AllocateDir.
@@ -887,6 +896,7 @@ PathNameOpenFile(FileName fileName, int fileFlags, int fileMode)
 	vfdP->fileFlags = fileFlags & ~(O_CREAT | O_TRUNC | O_EXCL);
 	vfdP->fileMode = fileMode;
 	vfdP->seekPos = 0;
+	vfdP->fileSize = 0;
 	vfdP->fdstate = 0x0;
 	vfdP->resowner = NULL;
 
@@ -1123,6 +1133,10 @@ FileClose(File file)
 			if (unlink(vfdP->fileName))
 				elog(LOG, "could not unlink file \"%s\": %m", vfdP->fileName);
 		}
+
+		/* Subtract its size from current usage */
+		temporary_files_size -= vfdP->fileSize;
+		vfdP->fileSize = 0;
 	}
 
 	/* Unregister it from the resource owner */
@@ -1242,6 +1256,31 @@ FileWrite(File file, char *buffer, int amount)
 	if (returnCode < 0)
 		return returnCode;
 
+	/*
+	 * If enforcing temp_file_limit and it's a temp file, check to see if the
+	 * write would overrun temp_file_limit, and throw error if so.  Note: it's
+	 * really a modularity violation to throw error here; we should set errno
+	 * and return -1.  However, there's no way to report a suitable error
+	 * message if we do that.  All current callers would just throw error
+	 * immediately anyway, so this is safe at present.
+	 */
+	if (temp_file_limit >= 0 && (VfdCache[file].fdstate & FD_TEMPORARY))
+	{
+		off_t	newPos = VfdCache[file].seekPos + amount;
+
+		if (newPos > VfdCache[file].fileSize)
+		{
+			uint64	newTotal = temporary_files_size;
+
+			newTotal += newPos - VfdCache[file].fileSize;
+			if (newTotal > (uint64) temp_file_limit * (uint64) 1024)
+				ereport(ERROR,
+						(errcode(ERRCODE_CONFIGURATION_LIMIT_EXCEEDED),
+						 errmsg("temporary file size exceeds temp_file_limit (%dkB)",
+								temp_file_limit)));
+		}
+	}
+
 retry:
 	errno = 0;
 	returnCode = write(VfdCache[file].fd, buffer, amount);
@@ -1251,7 +1290,21 @@ FileWrite(File file, char *buffer, int amount)
 		errno = ENOSPC;
 
 	if (returnCode >= 0)
+	{
 		VfdCache[file].seekPos += returnCode;
+
+		/* maintain fileSize and temporary_files_size if it's a temp file */
+		if (VfdCache[file].fdstate & FD_TEMPORARY)
+		{
+			off_t	newPos = VfdCache[file].seekPos;
+
+			if (newPos > VfdCache[file].fileSize)
+			{
+				temporary_files_size += newPos - VfdCache[file].fileSize;
+				VfdCache[file].fileSize = newPos;
+			}
+		}
+	}
 	else
 	{
 		/*
@@ -1854,11 +1907,11 @@ CleanupTempFiles(bool isProcExit)
 				if (fdstate & FD_TEMPORARY)
 				{
 					/*
-					 * If we're in the process of exiting a backend process, close
-					 * all temporary files. Otherwise, only close temporary files
-					 * local to the current transaction. They should be closed by
-					 * the ResourceOwner mechanism already, so this is just a
-					 * debugging cross-check.
+					 * If we're in the process of exiting a backend process,
+					 * close all temporary files. Otherwise, only close
+					 * temporary files local to the current transaction.
+					 * They should be closed by the ResourceOwner mechanism
+					 * already, so this is just a debugging cross-check.
 					 */
 					if (isProcExit)
 						FileClose(i);

src/backend/utils/errcodes.txt

@@ -4,7 +4,7 @@
 #
 # Copyright (c) 2003-2011, PostgreSQL Global Development Group
 #
-# This list serves a basis for generating source files containing error
+# This list serves as the basis for generating source files containing error
 # codes. It is kept in a common format to make sure all these source files have
 # the same contents.
 # The files generated from this one are:
@@ -24,14 +24,14 @@
 #      sqlstate    E/W/S    errcode_macro_name    spec_name
 #
 # where sqlstate is a five-character string following the SQLSTATE conventions,
-# the second field determines if the code means an error, a warning or success,
+# the second field indicates if the code means an error, a warning or success,
 # errcode_macro_name is the C macro name starting with ERRCODE that will be put
-# in errcodes.h and spec_name is a lowercase, underscore-separated name that
+# in errcodes.h, and spec_name is a lowercase, underscore-separated name that
 # will be used as the PL/pgSQL condition name and will also be included in the
 # SGML list. The last field is optional, if not present the PL/pgSQL condition
 # and the SGML entry will not be generated.
 #
-# Empty lines and ones starting with a hash are comments.
+# Empty lines and lines starting with a hash are comments.
 #
 # There are also special lines in the format of:
 #
@@ -368,6 +368,7 @@ Section: Class 53 - Insufficient Resources
 53100    E    ERRCODE_DISK_FULL                                              disk_full
 53200    E    ERRCODE_OUT_OF_MEMORY                                          out_of_memory
 53300    E    ERRCODE_TOO_MANY_CONNECTIONS                                   too_many_connections
+53400    E    ERRCODE_CONFIGURATION_LIMIT_EXCEEDED                           configuration_limit_exceeded
 
 Section: Class 54 - Program Limit Exceeded
 
@@ -393,7 +394,7 @@ Section: Class 57 - Operator Intervention
 57P01    E    ERRCODE_ADMIN_SHUTDOWN                                         admin_shutdown
 57P02    E    ERRCODE_CRASH_SHUTDOWN                                         crash_shutdown
 57P03    E    ERRCODE_CANNOT_CONNECT_NOW                                     cannot_connect_now
-57P04    E    ERRCODE_DATABASE_DROPPED										 database_dropped
+57P04    E    ERRCODE_DATABASE_DROPPED                                       database_dropped
 
 Section: Class 58 - System Error (errors external to PostgreSQL itself)
 

src/backend/utils/misc/guc.c

@@ -425,6 +425,8 @@ int			log_min_duration_statement = -1;
 int			log_temp_files = -1;
 int			trace_recovery_messages = LOG;
 
+int			temp_file_limit = -1;
+
 int			num_temp_buffers = 1024;
 
 char	   *data_directory;
@@ -535,6 +537,8 @@ const char *const config_group_names[] =
 	gettext_noop("Resource Usage"),
 	/* RESOURCES_MEM */
 	gettext_noop("Resource Usage / Memory"),
+	/* RESOURCES_DISK */
+	gettext_noop("Resource Usage / Disk"),
 	/* RESOURCES_KERNEL */
 	gettext_noop("Resource Usage / Kernel Resources"),
 	/* RESOURCES_VACUUM_DELAY */
@@ -1693,6 +1697,17 @@ static struct config_int ConfigureNamesInt[] =
 		check_max_stack_depth, assign_max_stack_depth, NULL
 	},
 
+	{
+		{"temp_file_limit", PGC_SUSET, RESOURCES_DISK,
+			gettext_noop("Limits the total size of all temp files used by each session."),
+			gettext_noop("-1 means no limit."),
+			GUC_UNIT_KB
+		},
+		&temp_file_limit,
+		-1, -1, INT_MAX,
+		NULL, NULL, NULL
+	},
+
 	{
 		{"vacuum_cost_page_hit", PGC_USERSET, RESOURCES_VACUUM_DELAY,
 			gettext_noop("Vacuum cost for a page found in the buffer cache."),

src/backend/utils/misc/postgresql.conf.sample

@@ -119,6 +119,11 @@
 #maintenance_work_mem = 16MB		# min 1MB
 #max_stack_depth = 2MB			# min 100kB
 
+# - Disk -
+
+#temp_file_limit = -1			# limits per-session temp file space
+					# in kB, or -1 for no limit
+
 # - Kernel Resource Usage -
 
 #max_files_per_process = 1000		# min 25

src/include/utils/guc.h

@@ -209,6 +209,8 @@ extern int	client_min_messages;
 extern int	log_min_duration_statement;
 extern int	log_temp_files;
 
+extern int	temp_file_limit;
+
 extern int	num_temp_buffers;
 
 extern char *data_directory;

src/include/utils/guc_tables.h

@@ -59,6 +59,7 @@ enum config_group
 	CONN_AUTH_SECURITY,
 	RESOURCES,
 	RESOURCES_MEM,
+	RESOURCES_DISK,
 	RESOURCES_KERNEL,
 	RESOURCES_VACUUM_DELAY,
 	RESOURCES_BGWRITER,