Fix archive_cleanup_command.

When I moved ExecuteRecoveryCommand() from xlog.c to xlogarchive.c, I didn't
realize that it's called from the checkpoint process, not the startup
process. I tried to use InRedo variable to decide whether or not to attempt
cleaning up the archive (must not do so before we have read the initial
checkpoint record), but that variable is only valid within the startup
process.

Instead, let ExecuteRecoveryCommand() always clean up the archive, and add
an explicit argument to RestoreArchivedFile() to say whether that's allowed
or not. The caller knows better.

Reported by Erik Rijkers, diagnosis by Fujii Masao. Only 9.3devel is
affected.

Author: hlinnaka

src/backend/access/transam/timeline.c

@@ -64,7 +64,7 @@ readTimeLineHistory(TimeLineID targetTLI)
 	if (InArchiveRecovery)
 	{
 		TLHistoryFileName(histfname, targetTLI);
-		RestoreArchivedFile(path, histfname, "RECOVERYHISTORY", 0);
+		RestoreArchivedFile(path, histfname, "RECOVERYHISTORY", 0, false);
 	}
 	else
 		TLHistoryFilePath(path, targetTLI);
@@ -153,7 +153,7 @@ existsTimeLineHistory(TimeLineID probeTLI)
 	if (InArchiveRecovery)
 	{
 		TLHistoryFileName(histfname, probeTLI);
-		RestoreArchivedFile(path, histfname, "RECOVERYHISTORY", 0);
+		RestoreArchivedFile(path, histfname, "RECOVERYHISTORY", 0, false);
 	}
 	else
 		TLHistoryFilePath(path, probeTLI);
@@ -257,7 +257,7 @@ writeTimeLineHistory(TimeLineID newTLI, TimeLineID parentTLI,
 	if (InArchiveRecovery)
 	{
 		TLHistoryFileName(histfname, parentTLI);
-		RestoreArchivedFile(path, histfname, "RECOVERYHISTORY", 0);
+		RestoreArchivedFile(path, histfname, "RECOVERYHISTORY", 0, false);
 	}
 	else
 		TLHistoryFilePath(path, parentTLI);

src/backend/access/transam/xlog.c

@@ -2588,7 +2588,8 @@ XLogFileRead(XLogSegNo segno, int emode, TimeLineID tli,
 
 			restoredFromArchive = RestoreArchivedFile(path, xlogfname,
 													  "RECOVERYXLOG",
-													  XLogSegSize);
+													  XLogSegSize,
+													  InRedo);
 			if (!restoredFromArchive)
 				return -1;
 			break;
@@ -9051,33 +9052,16 @@ GetXLogWriteRecPtr(void)
 }
 
 /*
- * Returns the redo pointer of the last restartpoint. This is the oldest
- * point in WAL that we still need, if we have to restart recovery. Returns
- * InvalidXLogRecPtr if we don't reliably know that point yet, that is,
- * before we have started WAL redo.
- *
- * This function only works in the startup process, and only while we are
- * in WAL redo. It's important to not return a value before redo has started,
- * to avoid deleting WAL files that we might still need, but there's no
- * fundamental reason why this couldn't return a valid value after redo has
- * finished, or in other processes. This is enough for the current usage,
- * however.
+ * Returns the redo pointer of the last checkpoint or restartpoint. This is
+ * the oldest point in WAL that we still need, if we have to restart recovery.
  */
 void
 GetOldestRestartPoint(XLogRecPtr *oldrecptr, TimeLineID *oldtli)
 {
-	if (InRedo)
-	{
-		LWLockAcquire(ControlFileLock, LW_SHARED);
-		*oldrecptr = ControlFile->checkPointCopy.redo;
-		*oldtli = ControlFile->checkPointCopy.ThisTimeLineID;
-		LWLockRelease(ControlFileLock);
-	}
-	else
-	{
-		*oldrecptr = InvalidXLogRecPtr;
-		*oldtli = 0;
-	}
+	LWLockAcquire(ControlFileLock, LW_SHARED);
+	*oldrecptr = ControlFile->checkPointCopy.redo;
+	*oldtli = ControlFile->checkPointCopy.ThisTimeLineID;
+	LWLockRelease(ControlFileLock);
 }
 
 /*

src/backend/access/transam/xlogarchive.c

@@ -41,10 +41,15 @@
  * For fixed-size files, the caller may pass the expected size as an
  * additional crosscheck on successful recovery.  If the file size is not
  * known, set expectedSize = 0.
+ *
+ * When 'cleanupEnabled' is false, refrain from deleting any old WAL segments
+ * in the archive. This is used when fetching the initial checkpoint record,
+ * when we are not yet sure how far back we need the WAL.
  */
 bool
 RestoreArchivedFile(char *path, const char *xlogfname,
-					const char *recovername, off_t expectedSize)
+					const char *recovername, off_t expectedSize,
+					bool cleanupEnabled)
 {
 	char		xlogpath[MAXPGPATH];
 	char		xlogRestoreCmd[MAXPGPATH];
@@ -113,9 +118,10 @@ RestoreArchivedFile(char *path, const char *xlogfname,
 	 * replication. All files earlier than this point can be deleted from the
 	 * archive, though there is no requirement to do so.
 	 *
-	 * We initialise this with the filename of an InvalidXLogRecPtr, which
-	 * will prevent the deletion of any WAL files from the archive because of
-	 * the alphabetic sorting property of WAL filenames.
+	 * If cleanup is not enabled, initialise this with the filename of
+	 * InvalidXLogRecPtr, which will prevent the deletion of any WAL files
+	 * from the archive because of the alphabetic sorting property of WAL
+	 * filenames.
 	 *
 	 * Once we have successfully located the redo pointer of the checkpoint
 	 * from which we start recovery we never request a file prior to the redo
@@ -124,9 +130,9 @@ RestoreArchivedFile(char *path, const char *xlogfname,
 	 * flags to signify the point when we can begin deleting WAL files from
 	 * the archive.
 	 */
-	GetOldestRestartPoint(&restartRedoPtr, &restartTli);
-	if (!XLogRecPtrIsInvalid(restartRedoPtr))
+	if (cleanupEnabled)
 	{
+		GetOldestRestartPoint(&restartRedoPtr, &restartTli);
 		XLByteToSeg(restartRedoPtr, restartSegNo);
 		XLogFileName(lastRestartPointFname, restartTli, restartSegNo);
 		/* we shouldn't need anything earlier than last restart point */

src/include/access/xlog_internal.h

@@ -243,7 +243,8 @@ extern char *recoveryRestoreCommand;
  * Prototypes for functions in xlogarchive.c
  */
 extern bool RestoreArchivedFile(char *path, const char *xlogfname,
-					const char *recovername, off_t expectedSize);
+					const char *recovername, off_t expectedSize,
+					bool cleanupEnabled);
 extern void ExecuteRecoveryCommand(char *command, char *commandName,
 					   bool failOnerror);
 extern void XLogArchiveNotify(const char *xlog);