Add pg_upgrade to /contrib; will be in 9.0 beta2.

Add documentation.

Supports migration from PG 8.3 and 8.4.

1 files changed, 228 insertions, 0 deletions

diff --git a/contrib/pg_upgrade/relfilenode.c b/contrib/pg_upgrade/relfilenode.c
new file mode 100644
index 00000000000..918447fcfaf
--- /dev/null
+++ b/contrib/pg_upgrade/relfilenode.c
@@ -0,0 +1,228 @@
+/*
+ *	relfilenode.c
+ *
+ *	relfilenode functions
+ */
+
+#include "pg_upgrade.h"
+
+#ifdef EDB_NATIVE_LANG
+#include <fcntl.h>
+#endif
+
+#include "catalog/pg_class.h"
+#include "access/transam.h"
+
+
+static void transfer_single_new_db(migratorContext *ctx, pageCnvCtx *pageConverter,
+					   FileNameMap *maps, int size);
+static void transfer_relfile(migratorContext *ctx, pageCnvCtx *pageConverter,
+				 const char *fromfile, const char *tofile,
+				 const char *oldnspname, const char *oldrelname,
+				 const char *newnspname, const char *newrelname);
+
+/*
+ * transfer_all_new_dbs()
+ *
+ * Responsible for upgrading all database. invokes routines to generate mappings and then
+ * physically link the databases.
+ */
+const char *
+transfer_all_new_dbs(migratorContext *ctx, DbInfoArr *olddb_arr,
+					 DbInfoArr *newdb_arr, char *old_pgdata, char *new_pgdata)
+{
+	int			dbnum;
+	const char *msg = NULL;
+
+	prep_status(ctx, "Restoring user relation files\n");
+
+	for (dbnum = 0; dbnum < newdb_arr->ndbs; dbnum++)
+	{
+		DbInfo	   *new_db = &newdb_arr->dbs[dbnum];
+		DbInfo	   *old_db = dbarr_lookup_db(olddb_arr, new_db->db_name);
+		FileNameMap *mappings;
+		int			n_maps;
+		pageCnvCtx *pageConverter = NULL;
+
+		n_maps = 0;
+		mappings = gen_db_file_maps(ctx, old_db, new_db, &n_maps, old_pgdata,
+									new_pgdata);
+
+		if (n_maps)
+		{
+			print_maps(ctx, mappings, n_maps, new_db->db_name);
+
+#ifdef PAGE_CONVERSION
+			msg = setupPageConverter(ctx, &pageConverter);
+#endif
+			transfer_single_new_db(ctx, pageConverter, mappings, n_maps);
+
+			pg_free(mappings);
+		}
+	}
+
+	prep_status(ctx, "");		/* in case nothing printed */
+	check_ok(ctx);
+
+	return msg;
+}
+
+
+/*
+ * get_pg_database_relfilenode()
+ *
+ *	Retrieves the relfilenode for a few system-catalog tables.	We need these
+ *	relfilenodes later in the upgrade process.
+ */
+void
+get_pg_database_relfilenode(migratorContext *ctx, Cluster whichCluster)
+{
+	PGconn	   *conn = connectToServer(ctx, "template1", whichCluster);
+	PGresult   *res;
+	int			i_relfile;
+
+	res = executeQueryOrDie(ctx, conn,
+							"SELECT c.relname, c.relfilenode "
+							"FROM 	pg_catalog.pg_class c, "
+							"		pg_catalog.pg_namespace n "
+							"WHERE 	c.relnamespace = n.oid AND "
+							"		n.nspname = 'pg_catalog' AND "
+							"		c.relname = 'pg_database' "
+							"ORDER BY c.relname");
+
+	i_relfile = PQfnumber(res, "relfilenode");
+	if (whichCluster == CLUSTER_OLD)
+		ctx->old.pg_database_oid = atol(PQgetvalue(res, 0, i_relfile));
+	else
+		ctx->new.pg_database_oid = atol(PQgetvalue(res, 0, i_relfile));
+
+	PQclear(res);
+	PQfinish(conn);
+}
+
+
+/*
+ * transfer_single_new_db()
+ *
+ * create links for mappings stored in "maps" array.
+ */
+static void
+transfer_single_new_db(migratorContext *ctx, pageCnvCtx *pageConverter,
+					   FileNameMap *maps, int size)
+{
+	int			mapnum;
+
+	for (mapnum = 0; mapnum < size; mapnum++)
+	{
+		char		old_file[MAXPGPATH];
+		char		new_file[MAXPGPATH];
+		struct dirent **namelist = NULL;
+		int			numFiles;
+
+		/* Copying files might take some time, so give feedback. */
+
+		snprintf(old_file, sizeof(old_file), "%s/%u", maps[mapnum].old_file, maps[mapnum].old);
+		snprintf(new_file, sizeof(new_file), "%s/%u", maps[mapnum].new_file, maps[mapnum].new);
+		pg_log(ctx, PG_REPORT, OVERWRITE_MESSAGE, old_file);
+
+		/*
+		 * Copy/link the relation file to the new cluster
+		 */
+		unlink(new_file);
+		transfer_relfile(ctx, pageConverter, old_file, new_file,
+						 maps[mapnum].old_nspname, maps[mapnum].old_relname,
+						 maps[mapnum].new_nspname, maps[mapnum].new_relname);
+
+		/* fsm/vm files added in PG 8.4 */
+		if (GET_MAJOR_VERSION(ctx->old.major_version) >= 804)
+		{
+			/*
+			 * Now copy/link any fsm and vm files, if they exist
+			 */
+			snprintf(scandir_file_pattern, sizeof(scandir_file_pattern), "%u_", maps[mapnum].old);
+			numFiles = pg_scandir(ctx, maps[mapnum].old_file, &namelist, dir_matching_filenames, NULL);
+
+			while (numFiles--)
+			{
+				snprintf(old_file, sizeof(old_file), "%s/%s", maps[mapnum].old_file,
+						 namelist[numFiles]->d_name);
+				snprintf(new_file, sizeof(new_file), "%s/%u%s", maps[mapnum].new_file,
+				  maps[mapnum].new, strchr(namelist[numFiles]->d_name, '_'));
+
+				unlink(new_file);
+				transfer_relfile(ctx, pageConverter, old_file, new_file,
+						  maps[mapnum].old_nspname, maps[mapnum].old_relname,
+						 maps[mapnum].new_nspname, maps[mapnum].new_relname);
+
+				pg_free(namelist[numFiles]);
+			}
+
+			pg_free(namelist);
+		}
+
+		/*
+		 * Now copy/link any related segments as well. Remember, PG breaks
+		 * large files into 1GB segments, the first segment has no extension,
+		 * subsequent segments are named relfilenode.1, relfilenode.2,
+		 * relfilenode.3, ...  'fsm' and 'vm' files use underscores so are not
+		 * copied.
+		 */
+		snprintf(scandir_file_pattern, sizeof(scandir_file_pattern), "%u.", maps[mapnum].old);
+		numFiles = pg_scandir(ctx, maps[mapnum].old_file, &namelist, dir_matching_filenames, NULL);
+
+		while (numFiles--)
+		{
+			snprintf(old_file, sizeof(old_file), "%s/%s", maps[mapnum].old_file,
+					 namelist[numFiles]->d_name);
+			snprintf(new_file, sizeof(new_file), "%s/%u%s", maps[mapnum].new_file,
+				  maps[mapnum].new, strchr(namelist[numFiles]->d_name, '.'));
+
+			unlink(new_file);
+			transfer_relfile(ctx, pageConverter, old_file, new_file,
+						  maps[mapnum].old_nspname, maps[mapnum].old_relname,
+						 maps[mapnum].new_nspname, maps[mapnum].new_relname);
+
+			pg_free(namelist[numFiles]);
+		}
+
+		pg_free(namelist);
+	}
+}
+
+
+/*
+ * transfer_relfile()
+ *
+ * Copy or link file from old cluster to new one.
+ */
+static void
+transfer_relfile(migratorContext *ctx, pageCnvCtx *pageConverter, const char *oldfile,
+		 const char *newfile, const char *oldnspname, const char *oldrelname,
+				 const char *newnspname, const char *newrelname)
+{
+	const char *msg;
+
+	if ((ctx->transfer_mode == TRANSFER_MODE_LINK) && (pageConverter != NULL))
+		pg_log(ctx, PG_FATAL, "this migration requires page-by-page conversion, "
+			   "you must use copy-mode instead of link-mode\n");
+
+	if (ctx->transfer_mode == TRANSFER_MODE_COPY)
+	{
+		pg_log(ctx, PG_INFO, "copying %s to %s\n", oldfile, newfile);
+
+		if ((msg = copyAndUpdateFile(ctx, pageConverter, oldfile, newfile, true)) != NULL)
+			pg_log(ctx, PG_FATAL, "error while copying %s.%s(%s) to %s.%s(%s): %s\n",
+				   oldnspname, oldrelname, oldfile, newnspname, newrelname, newfile, msg);
+	}
+	else
+	{
+		pg_log(ctx, PG_INFO, "linking %s to %s\n", newfile, oldfile);
+
+		if ((msg = linkAndUpdateFile(ctx, pageConverter, oldfile, newfile)) != NULL)
+			pg_log(ctx, PG_FATAL,
+			   "error while creating link from %s.%s(%s) to %s.%s(%s): %s\n",
+				   oldnspname, oldrelname, oldfile, newnspname, newrelname,
+				   newfile, msg);
+	}
+	return;
+}