Make WAL-related utilities handle .partial WAL files properly.

Commit de76884 changed an archive recovery so that the last WAL
segment with old timeline was renamed with suffix .partial. It should
have updated WAL-related utilities so that they can handle such
.paritial WAL files, but we forgot that.

This patch changes pg_archivecleanup so that it can clean up even
archived WAL files with .partial suffix. Also it allows us to specify
.partial WAL file name as the command-line argument "oldestkeptwalfile".

This patch also changes pg_resetxlog so that it can remove .partial
WAL files in pg_xlog directory.

pg_xlogdump cannot handle .partial WAL files. Per discussion,
we decided only to document that limitation instead of adding the fix.
Because a user can easily work around the limitation (i.e., just remove
.partial suffix from the file name) and the fix seems complicated for
very narrow use case.

Back-patch to 9.5 where the problem existed.

Review by Michael Paquier.
Discussion: http://www.postgresql.org/message-id/CAHGQGwGxMKnVHGgTfiig2Bt_2djec0in3-DLJmtg7+nEiidFdQ@mail.gmail.com

4 files changed, 43 insertions, 10 deletions

diff --git a/doc/src/sgml/ref/pg_xlogdump.sgml b/doc/src/sgml/ref/pg_xlogdump.sgml
index d9f4a6a499..1d78cf1a75 100644
--- a/doc/src/sgml/ref/pg_xlogdump.sgml
+++ b/doc/src/sgml/ref/pg_xlogdump.sgml
@@ -215,6 +215,12 @@ PostgreSQL documentation
     Only the specified timeline is displayed (or the default, if none is
     specified). Records in other timelines are ignored.
   </para>
+
+  <para>
+    <application>pg_xlogdump</> cannot read WAL files with suffix
+    <literal>.partial</>. If those files need to be read, <literal>.partial</>
+    suffix needs to be removed from the filename.
+  </para>
  </refsect1>
 
  <refsect1>
diff --git a/doc/src/sgml/ref/pgarchivecleanup.sgml b/doc/src/sgml/ref/pgarchivecleanup.sgml
index 779159d7fc..db39deaca1 100644
--- a/doc/src/sgml/ref/pgarchivecleanup.sgml
+++ b/doc/src/sgml/ref/pgarchivecleanup.sgml
@@ -60,8 +60,10 @@ archive_cleanup_command = 'pg_archivecleanup <replaceable>archivelocation</> %r'
   <para>
    When used as a standalone program all WAL files logically preceding the
    <replaceable>oldestkeptwalfile</> will be removed from <replaceable>archivelocation</>.
-   In this mode, if you specify a <filename>.backup</> file name, then only the file prefix
-   will be used as the <replaceable>oldestkeptwalfile</>. This allows you to remove
+   In this mode, if you specify a <filename>.partial</> or <filename>.backup</>
+   file name, then only the file prefix will be used as the
+   <replaceable>oldestkeptwalfile</>. This treatment of <filename>.backup</>
+   file name allows you to remove
    all WAL files archived prior to a specific base backup without error.
    For example, the following example will remove all files older than
    WAL file name <filename>000000010000003700000010</>:
diff --git a/src/bin/pg_archivecleanup/pg_archivecleanup.c b/src/bin/pg_archivecleanup/pg_archivecleanup.c
index 579a9bb843..c5569f32a3 100644
--- a/src/bin/pg_archivecleanup/pg_archivecleanup.c
+++ b/src/bin/pg_archivecleanup/pg_archivecleanup.c
@@ -125,7 +125,7 @@ CleanupPriorWALFiles(void)
 			 * file. Note that this means files are not removed in the order
 			 * they were originally written, in case this worries you.
 			 */
-			if (IsXLogFileName(walfile) &&
+			if ((IsXLogFileName(walfile) || IsPartialXLogFileName(walfile)) &&
 				strcmp(walfile + 8, exclusiveCleanupFileName + 8) < 0)
 			{
 				/*
@@ -181,7 +181,7 @@ CleanupPriorWALFiles(void)
  * SetWALFileNameForCleanup()
  *
  *	  Set the earliest WAL filename that we want to keep on the archive
- *	  and decide whether we need_cleanup
+ *	  and decide whether we need cleanup
  */
 static void
 SetWALFileNameForCleanup(void)
@@ -192,9 +192,10 @@ SetWALFileNameForCleanup(void)
 
 	/*
 	 * If restartWALFileName is a WAL file name then just use it directly. If
-	 * restartWALFileName is a .backup filename, make sure we use the prefix
-	 * of the filename, otherwise we will remove wrong files since
-	 * 000000010000000000000010.00000020.backup is after
+	 * restartWALFileName is a .partial or .backup filename, make sure we use
+	 * the prefix of the filename, otherwise we will remove wrong files since
+	 * 000000010000000000000010.partial and
+	 * 000000010000000000000010.00000020.backup are after
 	 * 000000010000000000000010.
 	 */
 	if (IsXLogFileName(restartWALFileName))
@@ -202,6 +203,26 @@ SetWALFileNameForCleanup(void)
 		strcpy(exclusiveCleanupFileName, restartWALFileName);
 		fnameOK = true;
 	}
+	else if (IsPartialXLogFileName(restartWALFileName))
+	{
+		int			args;
+		uint32		tli = 1,
+					log = 0,
+					seg = 0;
+
+		args = sscanf(restartWALFileName, "%08X%08X%08X.partial",
+					  &tli, &log, &seg);
+		if (args == 3)
+		{
+			fnameOK = true;
+
+			/*
+			 * Use just the prefix of the filename, ignore everything after
+			 * first period
+			 */
+			XLogFileNameById(exclusiveCleanupFileName, tli, log, seg);
+		}
+	}
 	else if (IsBackupHistoryFileName(restartWALFileName))
 	{
 		int			args;
diff --git a/src/bin/pg_resetxlog/pg_resetxlog.c b/src/bin/pg_resetxlog/pg_resetxlog.c
index e19a72b4c1..e7e8059a38 100644
--- a/src/bin/pg_resetxlog/pg_resetxlog.c
+++ b/src/bin/pg_resetxlog/pg_resetxlog.c
@@ -906,7 +906,8 @@ FindEndOfXLOG(void)
 
 	while (errno = 0, (xlde = readdir(xldir)) != NULL)
 	{
-		if (IsXLogFileName(xlde->d_name))
+		if (IsXLogFileName(xlde->d_name) ||
+			IsPartialXLogFileName(xlde->d_name))
 		{
 			unsigned int tli,
 						log,
@@ -976,7 +977,8 @@ KillExistingXLOG(void)
 
 	while (errno = 0, (xlde = readdir(xldir)) != NULL)
 	{
-		if (IsXLogFileName(xlde->d_name))
+		if (IsXLogFileName(xlde->d_name) ||
+			IsPartialXLogFileName(xlde->d_name))
 		{
 			snprintf(path, MAXPGPATH, "%s/%s", XLOGDIR, xlde->d_name);
 			if (unlink(path) < 0)
@@ -1028,7 +1030,9 @@ KillExistingArchiveStatus(void)
 	{
 		if (strspn(xlde->d_name, "0123456789ABCDEF") == XLOG_FNAME_LEN &&
 			(strcmp(xlde->d_name + XLOG_FNAME_LEN, ".ready") == 0 ||
-			 strcmp(xlde->d_name + XLOG_FNAME_LEN, ".done") == 0))
+			 strcmp(xlde->d_name + XLOG_FNAME_LEN, ".done") == 0 ||
+			 strcmp(xlde->d_name + XLOG_FNAME_LEN, ".partial.ready") == 0 ||
+			 strcmp(xlde->d_name + XLOG_FNAME_LEN, ".partial.done") == 0))
 		{
 			snprintf(path, MAXPGPATH, "%s/%s", ARCHSTATDIR, xlde->d_name);
 			if (unlink(path) < 0)