postgres.git / summary / log / commit / refs

commit    46e9019bbce96c309d27d4b164bf9a2d0d8292eb
Author:   Fujii Masao <fujii@postgresql.org>
Date:     Mon Aug 03 14:04:41 2015 +0000

    Make recovery rename tablespace_map to *.old if backup_label is not present.
    
    If tablespace_map file is present without backup_label file, there is
    no use of such file.  There is no harm in retaining it, but it is better
    to get rid of the map file so that we don't have any redundant file
    in data directory and it will avoid any sort of confusion. It seems
    prudent though to just rename the file out of the way rather than
    delete it completely, also we ignore any error that occurs in rename
    operation as even if map file is present without backup_label file,
    it is harmless.
    
    Back-patch to 9.5 where tablespace_map file was introduced.
    
    Amit Kapila, reviewed by Robert Haas, Alvaro Herrera and me.


 src/backend/access/transam/xlog.c | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index f06b51dde37..68e33eb1a99 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -5912,6 +5912,7 @@ StartupXLOG(void)
 	XLogReaderState *xlogreader;
 	XLogPageReadPrivate private;
 	bool		fast_promoted = false;
+	struct stat st;
 
 	/*
 	 * Read control file and check XLOG status looks valid.
@@ -6138,6 +6139,33 @@ StartupXLOG(void)
 	}
 	else
 	{
+		/*
+		 * If tablespace_map file is present without backup_label file, there
+		 * is no use of such file.  There is no harm in retaining it, but it
+		 * is better to get rid of the map file so that we don't have any
+		 * redundant file in data directory and it will avoid any sort of
+		 * confusion.  It seems prudent though to just rename the file out
+		 * of the way rather than delete it completely, also we ignore any
+		 * error that occurs in rename operation as even if map file is
+		 * present without backup_label file, it is harmless.
+		 */
+		if (stat(TABLESPACE_MAP, &st) == 0)
+		{
+			unlink(TABLESPACE_MAP_OLD);
+			if (rename(TABLESPACE_MAP, TABLESPACE_MAP_OLD) == 0)
+				ereport(LOG,
+					(errmsg("ignoring \"%s\" file because no \"%s\" file exists",
+							TABLESPACE_MAP, BACKUP_LABEL_FILE),
+					 errdetail("\"%s\" was renamed to \"%s\".",
+							   TABLESPACE_MAP, TABLESPACE_MAP_OLD)));
+			else
+				ereport(LOG,
+						(errmsg("ignoring \"%s\" file because no \"%s\" file exists",
+								TABLESPACE_MAP, BACKUP_LABEL_FILE),
+						 errdetail("Could not rename file \"%s\" to \"%s\": %m.",
+								   TABLESPACE_MAP, TABLESPACE_MAP_OLD)));
+		}
+
 		/*
 		 * It's possible that archive recovery was requested, but we don't
 		 * know how far we need to replay the WAL before we reach consistency.


[parent: 615b69595525]