def _get_transform(server1, server2, object1, object2, options): """Get the transformation SQL statements This method generates the SQL statements to transform the destination object based on direction of the compare. server1[in] first server connection server2[in] second server connection object1 the first object in the compare in the form: (db.name) object2 the second object in the compare in the form: (db.name) options[in] a dictionary containing the options for the operation: (quiet, etc.) Returns tuple - (bool - same db name?, list of transformation statements) """ from mysql.utilities.common.database import Database from mysql.utilities.common.sql_transform import SQLTransformer obj_type = None direction = options.get("changes-for", "server1") # If there is no dot, we do not have the format 'db_name.obj_name' for # object1 and therefore must treat it as a database name. if object1.find(".") == -1: obj_type = "DATABASE" # We are working with databases so db and name need to be set # to the database name to tell the get_object_definition() method # to retrieve the database information. db1 = object1 db2 = object2 name1 = object1 name2 = object2 else: try: db1, name1 = object1.split(".") db2, name2 = object2.split(".") except: raise UtilError("Invalid object name arguments for _get_transform" "(): %s, %s." % (object1, object2)) db_1 = Database(server1, db1, options) db_2 = Database(server2, db2, options) if obj_type is None: obj_type = db_1.get_object_type(name1) transform_str = [] obj1 = db_1.get_object_definition(db1, name1, obj_type) obj2 = db_2.get_object_definition(db2, name2, obj_type) # Get the transformation based on direction. transform_str = [] same_db_name = True xform = SQLTransformer(db_1, db_2, obj1[0], obj2[0], obj_type, options.get("verbosity", 0)) differences = xform.transform_definition() if differences is not None and len(differences) > 0: transform_str.extend(differences) return transform_str
def check_objects(self, server, db, events=True): """Check number of objects. Creates a string containing the number of objects for a given database. server[in] Server object to query db[in] name of database to check Returns string """ from mysql.utilities.common.database import Database db_source = Database(server, db) db_source.init() res = db_source.get_db_objects("TABLE") str = "OBJECT COUNTS: tables = %s, " % (len(res)) res = db_source.get_db_objects("VIEW") str += "views = %s, " % (len(res)) res = db_source.get_db_objects("TRIGGER") str += "triggers = %s, " % (len(res)) res = db_source.get_db_objects("PROCEDURE") str += "procedures = %s, " % (len(res)) res = db_source.get_db_objects("FUNCTION") str += "functions = %s, " % (len(res)) if events: res = db_source.get_db_objects("EVENT") str += "events = %s \n" % (len(res)) return str
def check_read_permissions(server, db_list, options): """ Check user permissions on server for specified databases. This method checks if the user used to establish the connection to the server has read permissions to access the specified lists of databases. server[in] Server instance. db_list[in] List of databases to check. options[in] Dictionary with access options: skip_views True = no views processed skip_proc True = no procedures processed skip_func True = no functions processed skip_grants True = no grants processed skip_events True = no events processed Returns an UtilDBError error if the server user does not have read permissions to access all specified databases or if any of them does not exist. """ for db_name in db_list: source_db = Database(server, db_name) # Error if source database does not exist. if not source_db.exists(): raise UtilDBError("Source database does not exist - " "{0}".format(db_name), -1, db_name) # Check privileges to access database. source_db.check_read_access(server.user, server.host, options)
def _copy_objects(source, destination, db_list, options, show_message=True, do_create=True): """Copy objects for a list of databases This method loops through a list of databases copying the objects as controlled by the skip options. source[in] Server class instance for source destination[in] Server class instance for destination options[in] copy options show_message[in] if True, display copy message Default = True do_create[in] if True, execute create statement for database Default = True """ # Copy objects for db_name in db_list: if show_message: # Display copy message if not options.get('quiet', False): msg = "# Copying database %s " % db_name[0] if db_name[1]: msg += "renamed as %s" % (db_name[1]) print msg # Get a Database class instance db = Database(source, db_name[0], options) # Perform the copy db.init() db.copy_objects(db_name[1], options, destination, options.get("threads", False), do_create)
def check_read_permissions(server, db_list, options): """ Check user permissions on server for specified databases. This method checks if the user used to establish the connection to the server has read permissions to access the specified lists of databases. server[in] Server instance. db_list[in] List of databases to check. options[in] Dictionary with access options: skip_views True = no views processed skip_proc True = no procedures processed skip_func True = no functions processed skip_grants True = no grants processed skip_events True = no events processed Returns an UtilDBError error if the server user does not have read permissions to access all specified databases or if any of them does not exist. """ for db_name in db_list: source_db = Database(server, db_name, options) # Error if source database does not exist. if not source_db.exists(): raise UtilDBError( "Source database does not exist - " "{0}".format(db_name), -1, db_name) # Check privileges to access database. source_db.check_read_access(server.user, server.host, options)
def _check_auto_increment(source, db_list, options): """Check auto increment values for 0 If any tables are found to have 0 in the list of databases, the code prints a warning along with a sample statement that can be used should the user decide she needs it when she does the import. source[in] Source connection db_list[in] List of databases to export options[in[ Global option list """ for db in db_list: db_obj = Database(source, db, options) # print warning if any tables have 0 as auto_increment value if db_obj.check_auto_increment(): sql_mode = source.show_server_variable("sql_mode") sql_mode_str = "NO_AUTO_VALUE_ON_ZERO" if sql_mode[0]: sql_mode_str = sql_mode[0][1] if 'NO_AUTO_VALUE_ON_ZERO' not in sql_mode[0][1]: sql_mode_str = ("'{0}',NO_AUTO_VALUE_ON_ZERO" "".format(sql_mode_str)) print(_AUTO_INC_WARNING) print("# SET SQL_MODE = '{0}'\n#".format(sql_mode_str))
def object_diff(server1_val, server2_val, object1, object2, options, object_type=None): """diff the definition of two objects Find the difference among two object definitions. server1_val[in] a dictionary containing connection information for the first server including: (user, password, host, port, socket) server2_val[in] a dictionary containing connection information for the second server including: (user, password, host, port, socket) object1[in] the first object in the compare in the form: (db.name) object2[in] the second object in the compare in the form: (db.name) options[in] a dictionary containing the options for the operation: (quiet, verbosity, difftype) object_type[in] type of the objects to be compared (e.g., TABLE, PROCEDURE, etc.). By default None (not defined). Returns None = objects are the same, diff[] = tables differ """ server1, server2 = server_connect(server1_val, server2_val, object1, object2, options) # Get the object type if unknown considering that objects of different # types can be found with the same name. if not object_type: #Get object types of object1 regexp_obj = re.compile(REGEXP_QUALIFIED_OBJ_NAME) m_obj = regexp_obj.match(object1) db_name, obj_name = m_obj.groups() db = Database(server1, db_name, options) obj1_types = db.get_object_type(obj_name) if not obj1_types: raise UtilDBError("The object {0} does not exist.".format(object1)) # Get object types of object2 m_obj = regexp_obj.match(object2) db_name, obj_name = m_obj.groups() db = Database(server2, db_name, options) obj2_types = db.get_object_type(obj_name) if not obj2_types: raise UtilDBError("The object {0} does not exist.".format(object2)) # Merge types found for both objects obj_types = set(obj1_types + obj2_types) # Diff objects considering all types found result = [] for obj_type in obj_types: res = diff_objects(server1, server2, object1, object2, options, obj_type) if res: result.append(res) return result if len(result) > 0 else None else: # Diff objects of known type return diff_objects(server1, server2, object1, object2, options, object_type)
def _get_objects(server, database, options): """Get all objects from the database (except grants) server[in] connected server object database[in] database names options[in] global options Returns list - objects in database """ options["skip_grants"] = True # Tell db class to skip grants db_obj = Database(server, database, options) if not db_obj.exists(): raise UtilDBError("The database does not exist: {0}".format(database)) db_obj.init() db_objects = db_obj.objects db_objects.sort() return db_objects
def get_create_object(server, object_name, options): """Get the object's create statement. This method retrieves the object create statement from the database. server[in] server connection object_name[in] name of object in the form db.objectname options[in] options: verbosity, quiet Returns string : create statement or raise error if object or db not exist """ from mysql.utilities.common.database import Database verbosity = options.get("verbosity", 0) quiet = options.get("quiet", False) db_name, sep, obj_name = object_name.partition(".") object = [db_name] db = Database(server, object[0], options) # Error if atabase does not exist if not db.exists(): raise UtilDBError("The database does not exist: {0}".format(object[0])) if not obj_name: object.append(object[0]) obj_type = "DATABASE" else: object.append(obj_name) obj_type = db.get_object_type(object[1]) if obj_type is None: raise UtilDBError( "The object {0} does not exist.".format(object_name)) create_stmt = db.get_create_statement(object[0], object[1], obj_type) if verbosity > 0 and not quiet: print "\n# Definition for object {0}:".format(object_name) print create_stmt return create_stmt
def get_create_object(server, object_name, options): """Get the object's create statement. This method retrieves the object create statement from the database. server[in] server connection object_name[in] name of object in the form db.objectname options[in] options: verbosity, quiet Returns string : create statement or raise error if object or db not exist """ from mysql.utilities.common.database import Database verbosity = options.get("verbosity", 0) quiet = options.get("quiet", False) db_name, sep, obj_name = object_name.partition(".") object = [db_name] db = Database(server, object[0], options) # Error if atabase does not exist if not db.exists(): raise UtilDBError("The database does not exist: {0}".format(object[0])) if not obj_name: object.append(object[0]) obj_type = "DATABASE" else: object.append(obj_name) obj_type = db.get_object_type(object[1]) if obj_type is None: raise UtilDBError("The object {0} does not exist.". format(object_name)) create_stmt = db.get_create_statement(object[0], object[1], obj_type) if verbosity > 0 and not quiet: print "\n# Definition for object {0}:".format(object_name) print create_stmt return create_stmt
def _get_objects(server, database, options): """Get all objects from the database (except grants) server[in] connected server object database[in] database names options[in] global options Returns list - objects in database """ from mysql.utilities.common.database import Database options["skip_grants"] = True # Tell db class to skip grants db_obj = Database(server, database, options) if not db_obj.exists(): raise UtilDBError("The database does not exist: {0}".format(database)) db_obj.init() db_objects = db_obj.objects db_objects.sort() return db_objects
def get_create_object(server, object_name, options, object_type): """Get the object's create statement. This method retrieves the object create statement from the database. server[in] server connection object_name[in] name of object in the form db.objectname options[in] options: verbosity, quiet object_type[in] type of the specified object (e.g, TABLE, PROCEDURE, etc.). Returns string : create statement or raise error if object or db not exist """ verbosity = options.get("verbosity", 0) quiet = options.get("quiet", False) m_obj = re.match(REGEXP_QUALIFIED_OBJ_NAME, object_name) db_name, obj_name = m_obj.groups() obj = [db_name] db = Database(server, obj[0], options) # Error if database does not exist if not db.exists(): raise UtilDBError("The database does not exist: {0}".format(obj[0])) if not obj_name or object_type == 'DATABASE': obj.append(db_name) else: obj.append(obj_name) create_stmt = db.get_create_statement(obj[0], obj[1], object_type) if verbosity > 0 and not quiet: print "\n# Definition for object {0}:".format(object_name) print create_stmt return create_stmt
def _get_transform(server1, server2, object1, object2, options, object_type): """Get the transformation SQL statements This method generates the SQL statements to transform the destination object based on direction of the compare. server1[in] first server connection server2[in] second server connection object1 the first object in the compare in the form: (db.name) object2 the second object in the compare in the form: (db.name) options[in] a dictionary containing the options for the operation: (quiet, etc.) object_type[in] type of the objects to be compared (e.g., TABLE, PROCEDURE, etc.). Returns tuple - (bool - same db name?, list of transformation statements) """ try: m_obj1 = re.match(REGEXP_QUALIFIED_OBJ_NAME, object1) db1, name1 = m_obj1.groups() m_obj2 = re.match(REGEXP_QUALIFIED_OBJ_NAME, object2) db2, name2 = m_obj2.groups() except: raise UtilError("Invalid object name arguments for _get_transform" "(): %s, %s." % (object1, object2)) # If the second part of the object qualified name is None, then the format # is not 'db_name.obj_name' for object1 and therefore must treat it as a # database name. (supports backticks and the use of '.' (dots) in names.) if not name1 or object_type == 'DATABASE': # We are working with databases so db and name need to be set # to the database name to tell the get_object_definition() method # to retrieve the database information. name1 = db1 name2 = db2 db_1 = Database(server1, db1, options) db_2 = Database(server2, db2, options) obj1 = db_1.get_object_definition(db1, name1, object_type) obj2 = db_2.get_object_definition(db2, name2, object_type) # Get the transformation based on direction. transform_str = [] xform = SQLTransformer(db_1, db_2, obj1[0], obj2[0], object_type, options.get('verbosity', 0)) differences = xform.transform_definition() if differences and len(differences) > 0: transform_str.extend(differences) return transform_str
def exists(self, tbl_name=None): """Check to see if the table exists tbl_name[in] table name (db.table) (optional) If omitted, operation is performed on the class instance table name. return True = table exists, False = table does not exist """ db, table = (None, None) if tbl_name: db, table = Database.parse_object_name(tbl_name) else: db = self.db_name table = self.tbl_name res = self.server.exec_query("SELECT TABLE_NAME " + "FROM INFORMATION_SCHEMA.TABLES " + "WHERE TABLE_SCHEMA = '%s'" % db + " and TABLE_NAME = '%s'" % table) return (res is not None and len(res) >= 1)
def export_metadata(source, src_val, db_list, options): """Produce rows to be used to recreate objects in a database. This method retrieves the objects for each database listed in the form of CREATE (SQL) statements or in a tabular form to the file specified. The valid values for the format parameter are SQL, CSV, TSV, VERTICAL, or GRID. source[in] Server instance src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) options[in] a dictionary containing the options for the copy: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, debug, exclude_names, exclude_patterns) Returns bool True = success, False = error """ from mysql.utilities.common.database import Database from mysql.utilities.common.format import format_tabular_list from mysql.utilities.common.format import format_vertical_list format = options.get("format", "sql") no_headers = options.get("no_headers", False) column_type = options.get("display", "brief") skip_create = options.get("skip_create", False) quiet = options.get("quiet", False) skip_tables = options.get("skip_tables", False) skip_views = options.get("skip_views", False) skip_triggers = options.get("skip_triggers", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) if options.get("all", False): rows = source.get_all_databases() for row in rows: db_list.append(row[0]) # Check user permissions on source for all databases for db_name in db_list: source_db = Database(source, db_name) # Make a dictionary of the options access_options = { 'skip_views' : skip_views, 'skip_procs' : skip_procs, 'skip_funcs' : skip_funcs, 'skip_grants' : skip_grants, 'skip_events' : skip_events, } source_db.check_read_access(src_val["user"], src_val["host"], access_options) for db_name in db_list: # Get a Database class instance db = Database(source, db_name, options) # Error is source database does not exist if not db.exists(): raise UtilDBError("Source database does not exist - %s" % db_name, -1, db_name) if not quiet: print "# Exporting metadata from %s" % db_name # Perform the extraction if format == "sql": db.init() # quote database name with backticks q_db_name = quote_with_backticks(db_name) if not skip_create: print "DROP DATABASE IF EXISTS %s;" % q_db_name print "CREATE DATABASE %s;" % q_db_name print "USE %s;" % q_db_name for dbobj in db.get_next_object(): if dbobj[0] == "GRANT" and not skip_grants: if not quiet: print "# Grant:" if dbobj[1][3]: create_str = "GRANT %s ON %s.%s TO %s;" % \ (dbobj[1][1], q_db_name, quote_with_backticks(dbobj[1][3]), dbobj[1][0]) else: create_str = "GRANT %s ON %s.* TO %s;" % \ (dbobj[1][1], q_db_name, dbobj[1][0]) if create_str.find("%"): create_str = re.sub("%", "%%", create_str) print create_str else: if not quiet: print "# %s: %s.%s" % (dbobj[0], db_name, dbobj[1][0]) if (dbobj[0] == "PROCEDURE" and not skip_procs) or \ (dbobj[0] == "FUNCTION" and not skip_funcs) or \ (dbobj[0] == "EVENT" and not skip_events) or \ (dbobj[0] == "TRIGGER" and not skip_triggers): print "DELIMITER ||" print "%s;" % db.get_create_statement(db_name, dbobj[1][0], dbobj[0]) if (dbobj[0] == "PROCEDURE" and not skip_procs) or \ (dbobj[0] == "FUNCTION" and not skip_funcs) or \ (dbobj[0] == "EVENT" and not skip_events) or \ (dbobj[0] == "TRIGGER" and not skip_triggers): print "||" print "DELIMITER ;" else: objects = [] if not skip_tables: objects.append("TABLE") if not skip_views: objects.append("VIEW") if not skip_triggers: objects.append("TRIGGER") if not skip_procs: objects.append("PROCEDURE") if not skip_funcs: objects.append("FUNCTION") if not skip_events: objects.append("EVENT") if not skip_grants: objects.append("GRANT") for obj_type in objects: sys.stdout.write("# %sS in %s:" % (obj_type, db_name)) if format in ('grid', 'vertical'): rows = db.get_db_objects(obj_type, column_type, True) else: rows = db.get_db_objects(obj_type, column_type, True, True) if len(rows[1]) < 1: print " (none found)" else: print # Cannot use print_list here becasue we must manipulate # the behavior of format_tabular_list list_options = {} if format == "vertical": format_vertical_list(sys.stdout, rows[0], rows[1]) elif format == "tab": list_options['print_header'] = not no_headers list_options['separator'] = '\t' format_tabular_list(sys.stdout, rows[0], rows[1], list_options) elif format == "csv": list_options['print_header'] = not no_headers list_options['separator'] = ',' format_tabular_list(sys.stdout, rows[0], rows[1], list_options) else: # default to table format format_tabular_list(sys.stdout, rows[0], rows[1]) if not quiet: print "#...done." return True
def import_file(dest_val, file_name, options): """Import a file This method reads a file and, if needed, transforms the file into discrete SQL statements for execution on the destination server. It accepts any of the formal structured files produced by the mysqlexport utility including formats SQL, CSV, TAB, GRID, and VERTICAL. It will read these files and skip or include the definitions or data as specified in the options. An error is raised for any conversion errors or errors while executing the statements. Users are highly encouraged to use the --dryrun option which will print the SQL statements without executing them. dest_val[in] a dictionary containing connection information for the destination including: (user, password, host, port, socket) file_name[in] name (and path) of the file to import options[in] a dictionary containing the options for the import: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, and debug) Returns bool True = success, False = error """ from mysql.utilities.common.database import Database from mysql.utilities.common.options import check_engine_options from mysql.utilities.common.table import Table from mysql.utilities.common.server import connect_servers # Helper method to dig through the definitions for create statements def _process_definitions(statements, table_col_list, db_name): # First, get the SQL strings sql_strs = _build_create_objects(obj_type, db_name, definitions) statements.extend(sql_strs) # Now, save the column list col_list = _build_col_metadata(obj_type, definitions) if len(col_list) > 0: table_col_list.extend(col_list) def _process_data(tbl_name, statements, columns, table_col_list, table_rows, skip_blobs): # if there is data here, build bulk inserts # First, create table reference, then call insert_rows() tbl = Table(destination, tbl_name) # Need to check to see if table exists! if tbl.exists(): tbl.get_column_metadata() col_meta = True elif len(table_col_list) > 0: col_meta = _get_column_metadata(tbl, table_col_list) else: fix_cols = [] fix_cols.append((tbl.tbl_name, columns)) col_meta = _get_column_metadata(tbl, fix_cols) if not col_meta: raise UtilError("Cannot build bulk insert statements without " "the table definition.") ins_strs = tbl.make_bulk_insert(table_rows, tbl.q_db_name) if len(ins_strs[0]) > 0: statements.extend(ins_strs[0]) if len(ins_strs[1]) > 0 and not skip_blobs: for update in ins_strs[1]: statements.append(update) # Gather options format = options.get("format", "sql") no_headers = options.get("no_headers", False) quiet = options.get("quiet", False) import_type = options.get("import_type", "definitions") single = options.get("single", True) dryrun = options.get("dryrun", False) do_drop = options.get("do_drop", False) skip_blobs = options.get("skip_blobs", False) skip_gtid = options.get("skip_gtid", False) # Attempt to connect to the destination server conn_options = { 'quiet': quiet, 'version': "5.1.30", } servers = connect_servers(dest_val, None, conn_options) destination = servers[0] # Check storage engines check_engine_options(destination, options.get("new_engine", None), options.get("def_engine", None), False, options.get("quiet", False)) if not quiet: if import_type == "both": str = "definitions and data" else: str = import_type print "# Importing %s from %s." % (str, file_name) # Setup variables we will need skip_header = not no_headers if format == "sql": skip_header = False get_db = True check_privileges = False db_name = None file = open(file_name) columns = [] read_columns = False table_rows = [] obj_type = "" definitions = [] statements = [] table_col_list = [] tbl_name = "" skip_rpl = options.get("skip_rpl", False) gtid_command_found = False supports_gtid = servers[0].supports_gtid() == 'ON' skip_gtid_warning_printed = False gtid_version_checked = False # Read the file one object/definition group at a time for row in read_next(file, format): # Check for replication command if row[0] == "RPL_COMMAND": if not skip_rpl: statements.append(row[1]) continue if row[0] == "GTID_COMMAND": gtid_command_found = True if not supports_gtid: # only display warning once if not skip_gtid_warning_printed: print _GTID_SKIP_WARNING skip_gtid_warning_printed = True elif not skip_gtid: if not gtid_version_checked: gtid_version_checked = True # Check GTID version for complete feature support servers[0].check_gtid_version() # Check the gtid_purged value too servers[0].check_gtid_executed("import") statements.append(row[1]) continue # If this is the first pass, get the database name from the file if get_db: if skip_header: skip_header = False else: db_name = _get_db(row) # quote db_name with backticks if needed if db_name and not is_quoted_with_backticks(db_name): db_name = quote_with_backticks(db_name) get_db = False if do_drop and import_type != "data": statements.append("DROP DATABASE IF EXISTS %s;" % db_name) if import_type != "data": if not _skip_object("CREATE_DB", options) and \ not format == 'sql': statements.append("CREATE DATABASE %s;" % db_name) # This is the first time through the loop so we must # check user permissions on source for all databases if db_name is not None: dest_db = Database(destination, db_name) # Make a dictionary of the options access_options = options.copy() dest_db.check_write_access(dest_val['user'], dest_val['host'], access_options) # Now check to see if we want definitions, data, or both: if row[0] == "sql" or row[0] in _DEFINITION_LIST: if format != "sql" and len(row[1]) == 1: raise UtilError("Cannot read an import file generated with " "--display=NAMES") if import_type in ("definitions", "both"): if format == "sql": statements.append(row[1]) else: if obj_type == "": obj_type = row[0] if obj_type != row[0]: if len(definitions) > 0: _process_definitions(statements, table_col_list, db_name) obj_type = row[0] definitions = [] if not _skip_object(row[0], options): definitions.append(row[1]) else: # see if there are any definitions to process if len(definitions) > 0: _process_definitions(statements, table_col_list, db_name) definitions = [] if import_type in ("data", "both"): if _skip_object("DATA", options): continue # skip data elif format == "sql": statements.append(row[1]) else: if row[0] == "BEGIN_DATA": # Start of table so first row is columns. if len(table_rows) > 0: _process_data(tbl_name, statements, columns, table_col_list, table_rows, skip_blobs) table_rows = [] read_columns = True tbl_name = row[1] if not is_quoted_with_backticks(tbl_name): db, sep, tbl = tbl_name.partition('.') q_db = quote_with_backticks(db) q_tbl = quote_with_backticks(tbl) tbl_name = ".".join([q_db, q_tbl]) else: if read_columns: columns = row[1] read_columns = False else: if not single: table_rows.append(row[1]) else: str = _build_insert_data( columns, tbl_name, row[1]) statements.append(str) # Process remaining definitions if len(definitions) > 0: _process_definitions(statements, table_col_list, db_name) definitions = [] # Process remaining data rows if len(table_rows) > 0: _process_data(tbl_name, statements, columns, table_col_list, table_rows, skip_blobs) table_rows = [] # Now process the statements _exec_statements(statements, destination, format, options, dryrun) file.close() # Check gtid process if supports_gtid and not gtid_command_found: print _GTID_MISSING_WARNING if not quiet: print "#...done." return True
def _export_data(source, server_values, db_list, output_file, options): """Export data from the specified list of databases. This private method retrieves the data for each specified databases in SQL format (e.g., INSERT statements) or in a tabular form (GRID, TAB, CSV, VERTICAL) to the specified file. This private method does not check permissions. source[in] Server instance. server_values[in] Server connection values. db_list[in] List of databases to export. output_file[in] Output file to store the export data. options[in] Dictionary containing the options for the export: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, file_per_tbl, and debug). """ frmt = options.get("format", "sql") quiet = options.get("quiet", False) file_per_table = options.get("file_per_tbl", False) sql_mode = source.select_variable("SQL_MODE") # Get tables list. table_list = [] for db_name in db_list: source_db = Database(source, db_name, options) # Build table list. tables = source_db.get_db_objects("TABLE") for table in tables: table_list.append((db_name, table[0])) previous_db = "" export_tbl_tasks = [] for table in table_list: # Determine start for processing table from a different database. db_name = table[0] if previous_db != db_name: previous_db = db_name if not quiet: q_db_name = quote_with_backticks(db_name, sql_mode) if frmt == "sql": output_file.write("USE {0};\n".format(q_db_name)) output_file.write( "# Exporting data from {0}\n".format(q_db_name)) if file_per_table: output_file.write("# Writing table data to files.\n") # Print sample SOURCE command warning even in quiet mode. if file_per_table and frmt == 'sql': output_file.write("# The following are sample SOURCE commands." " If needed correct the path to match files " "location.\n") # Check multiprocess table export (only on POSIX systems). if options['multiprocess'] > 1 and os.name == 'posix': # Create export task. # Note: Server connection values are passed in the task dictionary # instead of a server instance, otherwise a multiprocessing error # is issued when assigning the task to a worker. export_task = { 'srv_con': server_values, 'table': table, 'options': options, } export_tbl_tasks.append(export_task) else: # Export data from a table (no multiprocessing). _export_table_data(source, table, output_file, options) # Print SOURCE command if --file-per-table is used and format is SQL. if file_per_table and frmt == 'sql': tbl_name = ".".join(table) output_file.write("# SOURCE {0}\n".format( _generate_tbl_filename(tbl_name, frmt))) # Export tables concurrently. if export_tbl_tasks: # Create process pool. workers_pool = multiprocessing.Pool(processes=options['multiprocess']) # Concurrently export tables. res = workers_pool.map_async(multiprocess_tbl_export_task, export_tbl_tasks) workers_pool.close() # Get list of temporary files with the exported data. tmp_files_list = res.get() workers_pool.join() # Merge resulting temp files (if generated). for tmp_filename in tmp_files_list: if tmp_filename: tmp_file = open(tmp_filename, 'r') shutil.copyfileobj(tmp_file, output_file) tmp_file.close() os.remove(tmp_filename) if not quiet: output_file.write("#...done.\n")
def check_index(src_val, table_args, options): """Check for duplicate or redundant indexes for one or more tables This method will examine the indexes for one or more tables and identify any indexes that are potential duplicates or redundant. It prints the equivalent DROP statements if selected. src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) table_args[in] list of tables in the form 'db.table' or 'db' options[in] dictionary of options to include: show-drops : show drop statements for dupe indexes skip : skip non-existant tables verbosity : print extra information show-indexes : show all indexes for each table index-format : index format = sql, table, tab, csv worst : show worst performing indexes best : show best performing indexes Returns bool True = success, raises UtilError if error """ # Get options show_drops = options.get("show-drops", False) skip = options.get("skip", False) verbosity = options.get("verbosity", False) show_indexes = options.get("show-indexes", False) index_format = options.get("index-format", False) stats = options.get("stats", False) first_indexes = options.get("best", None) last_indexes = options.get("worst", None) from mysql.utilities.common.server import connect_servers from mysql.utilities.common.database import Database from mysql.utilities.common.table import Table # Try to connect to the MySQL database server. conn_options = { 'quiet' : verbosity == 1, 'version' : "5.0.0", } servers = connect_servers(src_val, None, conn_options) source = servers[0] db_list = [] # list of databases table_list = [] # list of all tables to process # Build a list of objects to process # 1. start with db_list if no obects present on command line # 2. process command line options. # 3. loop through database list and add all tables # 4. check indexes # Perform the options check here. Loop through objects presented. for obj in table_args: # If a . appears, we are operating on a specific table idx = obj.count(".") if (idx == 1): table_list.append(obj) # Else we are operating on a specific database. else: db_list.append(obj) # Loop through database list adding tables for db in db_list: db_source = Database(source, db) db_source.init() tables = db_source.get_db_objects("TABLE") if not tables and verbosity >= 1: print "# Warning: database %s does not exist. Skipping." % (db) for table in tables: table_list.append(db + "." + table[0]) # Fail if no tables to check if not table_list: raise UtilError("No tables to check.") if verbosity > 1: print "# Checking indexes..." # Check indexes for each table in the list for table_name in table_list: tbl_options = { 'verbose' : verbosity >= 1, 'get_cols' : False, 'quiet' : verbosity is None or verbosity < 1 } tbl = Table(source, table_name, tbl_options) exists = tbl.exists() if not exists and not skip: raise UtilError("Table %s does not exist. Use --skip " "to skip missing tables." % table_name) if exists: if not tbl.get_indexes(): if verbosity > 1: print "# Table %s is not indexed." % (table_name) else: if show_indexes: tbl.print_indexes(index_format) # Show if table has primary key if not tbl.has_primary_key(): if verbosity > 1: print "# Table %s does not contain a PRIMARY key." tbl.check_indexes(show_drops) # Show best and/or worst indexes if stats: if first_indexes is not None: tbl.show_special_indexes(index_format, first_indexes, True) if last_indexes is not None: tbl.show_special_indexes(index_format, last_indexes) if verbosity > 1: print "#" if verbosity > 1: print "# ...done."
def get_copy_lock(server, db_list, options, include_mysql=False, cloning=False): """Get an instance of the Lock class with a standard copy (read) lock This method creates an instance of the Lock class using the lock type specified in the options. It is used to initiate the locks for the copy and related operations. server[in] Server instance for locking calls db_list[in] list of database names options[in] option dictionary Must include the skip_* options for copy and export include_mysql[in] if True, include the mysql tables for copy operation cloning[in] if True, create lock tables with WRITE on dest db Default = False Returns Lock - Lock class instance """ rpl_mode = options.get("rpl_mode", None) locking = options.get('locking', 'snapshot') # Determine if we need to use FTWRL. There are two conditions: # - running on master (rpl_mode = 'master') # - using locking = 'lock-all' and rpl_mode present if (rpl_mode in ["master", "both"]) or \ (rpl_mode and locking == 'lock-all'): new_opts = options.copy() new_opts['locking'] = 'flush' lock = Lock(server, [], new_opts) # if this is a lock-all type and not replication operation, # find all tables and lock them elif locking == 'lock-all': table_lock_list = [] # Build table lock list for db_name in db_list: db = db_name[0] if type(db_name) == tuple else db_name source_db = Database(server, db) tables = source_db.get_db_objects("TABLE") for table in tables: table_lock_list.append(("{0}.{1}".format(db, table[0]), 'READ')) # Cloning requires issuing WRITE locks because we use same # conn. # Non-cloning will issue WRITE lock on a new destination conn. if cloning: if db_name[1] is None: db_clone = db_name[0] else: db_clone = db_name[1] # For cloning, we use the same connection so we need to # lock the destination tables with WRITE. table_lock_list.append(("{0}.{1}".format(db_clone, table[0]), 'WRITE')) # We must include views for server version 5.5.3 and higher if server.check_version_compat(5, 5, 3): tables = source_db.get_db_objects("VIEW") for table in tables: table_lock_list.append(("{0}.{1}".format(db, table[0]), 'READ')) # Cloning requires issuing WRITE locks because we use same # conn. # Non-cloning will issue WRITE lock on a new destination # conn. if cloning: if db_name[1] is None: db_clone = db_name[0] else: db_clone = db_name[1] # For cloning, we use the same connection so we need to # lock the destination tables with WRITE. table_lock_list.append(("{0}.{1}".format(db_clone, table[0]), 'WRITE')) # Now add mysql tables if include_mysql: # Don't lock proc tables if no procs of funcs are being read if not options.get('skip_procs', False) and \ not options.get('skip_funcs', False): table_lock_list.append(("mysql.proc", 'READ')) table_lock_list.append(("mysql.procs_priv", 'READ')) # Don't lock event table if events are skipped if not options.get('skip_events', False): table_lock_list.append(("mysql.event", 'READ')) lock = Lock(server, table_lock_list, options) # Use default or no locking option else: lock = Lock(server, [], options) return lock
def _export_metadata(source, db_list, output_file, options): """Export metadata from the specified list of databases. This private method retrieves the objects metadata for each database listed in the form of CREATE (SQL) statements or in a tabular form (GRID, TAB, CSV, VERTICAL) to the specified file. This private method does not check permissions. source[in] Server instance. db_list[in] List of databases to export. output_file[in] Output file to store the metadata information. options[in] Dictionary containing the options for the export: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, debug, exclude_names, exclude_patterns) """ frmt = options.get("format", "sql") no_headers = options.get("no_headers", False) column_type = options.get("display", "brief") quiet = options.get("quiet", False) skip_create = options.get("skip_create", False) skip_tables = options.get("skip_tables", False) skip_views = options.get("skip_views", False) skip_triggers = options.get("skip_triggers", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) for db_name in db_list: # Get a Database class instance db = Database(source, db_name, options) # Export database metadata if not quiet: output_file.write( "# Exporting metadata from {0}\n".format(db.db_name) ) # Perform the extraction if frmt == "sql": db.init() if not skip_create: output_file.write( "DROP DATABASE IF EXISTS {0};\n".format(db.q_db_name) ) output_file.write( "CREATE DATABASE {0};\n".format(db.q_db_name) ) output_file.write("USE {0};\n".format(db.q_db_name)) for dbobj in db.get_next_object(): if dbobj[0] == "GRANT" and not skip_grants: if not quiet: output_file.write("# Grant:\n") if dbobj[1][3]: create_str = "GRANT {0} ON {1}.{2} TO {3};\n".format( dbobj[1][1], db.q_db_name, quote_with_backticks(dbobj[1][3]), dbobj[1][0] ) else: create_str = "GRANT {0} ON {1}.* TO {2};\n".format( dbobj[1][1], db.q_db_name, dbobj[1][0] ) output_file.write(create_str) else: if not quiet: output_file.write( "# {0}: {1}.{2}\n".format(dbobj[0], db.db_name, dbobj[1][0]) ) if (dbobj[0] == "PROCEDURE" and not skip_procs) or \ (dbobj[0] == "FUNCTION" and not skip_funcs) or \ (dbobj[0] == "EVENT" and not skip_events) or \ (dbobj[0] == "TRIGGER" and not skip_triggers): output_file.write("DELIMITER ||\n") output_file.write("{0};\n".format( db.get_create_statement(db.db_name, dbobj[1][0], dbobj[0]) )) if (dbobj[0] == "PROCEDURE" and not skip_procs) or \ (dbobj[0] == "FUNCTION" and not skip_funcs) or \ (dbobj[0] == "EVENT" and not skip_events) or \ (dbobj[0] == "TRIGGER" and not skip_triggers): output_file.write("||\n") output_file.write("DELIMITER ;\n") else: objects = [] if not skip_tables: objects.append("TABLE") if not skip_funcs: objects.append("FUNCTION") if not skip_procs: objects.append("PROCEDURE") if not skip_views: objects.append("VIEW") if not skip_triggers: objects.append("TRIGGER") if not skip_events: objects.append("EVENT") if not skip_grants: objects.append("GRANT") for obj_type in objects: output_file.write( "# {0}S in {1}:".format(obj_type, db.db_name) ) if frmt in ('grid', 'vertical'): rows = db.get_db_objects(obj_type, column_type, True) else: rows = db.get_db_objects(obj_type, column_type, True, True) if len(rows[1]) < 1: output_file.write(" (none found)\n") else: output_file.write("\n") # Cannot use print_list here because we must manipulate # the behavior of format_tabular_list. list_options = {} if frmt == "vertical": format_vertical_list(output_file, rows[0], rows[1]) elif frmt == "tab": list_options['print_header'] = not no_headers list_options['separator'] = '\t' format_tabular_list(output_file, rows[0], rows[1], list_options) elif frmt == "csv": list_options['print_header'] = not no_headers list_options['separator'] = ',' format_tabular_list(output_file, rows[0], rows[1], list_options) else: # default to table format format_tabular_list(output_file, rows[0], rows[1]) if not quiet: output_file.write("#...done.\n")
def object_diff(server1_val, server2_val, object1, object2, options, object_type=None): """diff the definition of two objects Find the difference among two object definitions. server1_val[in] a dictionary containing connection information for the first server including: (user, password, host, port, socket) server2_val[in] a dictionary containing connection information for the second server including: (user, password, host, port, socket) object1[in] the first object in the compare in the form: (db.name) object2[in] the second object in the compare in the form: (db.name) options[in] a dictionary containing the options for the operation: (quiet, verbosity, difftype) object_type[in] type of the objects to be compared (e.g., TABLE, PROCEDURE, etc.). By default None (not defined). Returns None = objects are the same, diff[] = tables differ """ if isinstance(server1_val, dict): # dict or common.server.Server object server1, server2 = server_connect(server1_val, server2_val, object1, object2, options) else: # to save connection server1, server2 = server1_val, server2_val force = options.get("force", None) # compare db's all objects include_create = options.get("include_create", False) # db1.*:db2.* if include_create and object1.endswith('.*') and object2.endswith('.*'): direction = options.get("changes-for", None) reverse = options.get("reverse", False) db_name1, _ = parse_object_name(object1, server1.select_variable("SQL_MODE")) db_name2, _ = parse_object_name(object2, server2.select_variable("SQL_MODE")) in_both, in_db1, in_db2 = get_common_objects(server1, server2, db_name1, db_name2, True, options) # create/alter/drop need all objects compare all_object = set(in_both + in_db1 + in_db2) # call myself recusively to compare all objects for this_obj in all_object: object1 = db_name1 + "." + this_obj[1][0] object2 = db_name2 + "." + this_obj[1][0] # share the same connection in this loop. object_type=None object_diff(server1, server2, object1, object2, options, object_type=None) return [] # Get the object type if unknown considering that objects of different # types can be found with the same name. if not object_type: # Get object types of object1 sql_mode = server1.select_variable("SQL_MODE") db_name, obj_name = parse_object_name(object1, sql_mode) db = Database(server1, db_name, options) obj1_types = db.get_object_type(obj_name) if not obj1_types: if include_create: # if allow generating create object ddl, give 'NULL' object here to tell common.dbcompare.py to handle obj1_types = ['NULL'] else: msg = "The object {0} does not exist.".format(object1) if not force: raise UtilDBError(msg) print("ERROR: {0}".format(msg)) return [] # Get object types of object2 sql_mode = server2.select_variable("SQL_MODE") db_name, obj_name = parse_object_name(object2, sql_mode) db = Database(server2, db_name, options) obj2_types = db.get_object_type(obj_name) if not obj2_types: if include_create: obj2_types = ['NULL'] else: msg = "The object {0} does not exist.".format(object2) if not force: raise UtilDBError(msg) print("ERROR: {0}".format(msg)) return [] # Merge types found for both objects obj_types = set(obj1_types + obj2_types) if obj_types == set(['NULL']): msg = "The object {0} or {1} does not exist in the source side.".format( object1, object2) if not force: raise UtilDBError(msg) print("ERROR: {0}".format(msg)) return [] elif 'NULL' in obj_types: # at least one object exist in db1 and db2 # new db object like TABLE-NULL or NULL-TABLE , 'TABLE' is needed for after use in diff_objects() obj_types = set(['-'.join(obj1_types + obj2_types)]) # Diff objects considering all types found result = [] for obj_type in obj_types: res = diff_objects(server1, server2, object1, object2, options, obj_type) if res: result.append(res) return result if len(result) > 0 else None else: # Diff objects of known type return diff_objects(server1, server2, object1, object2, options, object_type)
def check_index(src_val, table_args, options): """Check for duplicate or redundant indexes for one or more tables This method will examine the indexes for one or more tables and identify any indexes that are potential duplicates or redundant. It prints the equivalent DROP statements if selected. src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) table_args[in] list of tables in the form 'db.table' or 'db' options[in] dictionary of options to include: show-drops : show drop statements for dupe indexes skip : skip non-existant tables verbosity : print extra information show-indexes : show all indexes for each table index-format : index format = sql, table, tab, csv worst : show worst performing indexes best : show best performing indexes Returns bool True = success, raises UtilError if error """ # Get options show_drops = options.get("show-drops", False) skip = options.get("skip", False) verbosity = options.get("verbosity", False) show_indexes = options.get("show-indexes", False) index_format = options.get("index-format", False) stats = options.get("stats", False) first_indexes = options.get("best", None) last_indexes = options.get("worst", None) from mysql.utilities.common.server import connect_servers from mysql.utilities.common.database import Database from mysql.utilities.common.table import Table # Try to connect to the MySQL database server. conn_options = { 'quiet': verbosity == 1, 'version': "5.0.0", } servers = connect_servers(src_val, None, conn_options) source = servers[0] db_list = [] # list of databases table_list = [] # list of all tables to process # Build a list of objects to process # 1. start with db_list if no obects present on command line # 2. process command line options. # 3. loop through database list and add all tables # 4. check indexes # Perform the options check here. Loop through objects presented. for obj in table_args: # If a . appears, we are operating on a specific table idx = obj.count(".") if (idx == 1): table_list.append(obj) # Else we are operating on a specific database. else: db_list.append(obj) # Loop through database list adding tables for db in db_list: db_source = Database(source, db) db_source.init() tables = db_source.get_db_objects("TABLE") if not tables and verbosity >= 1: print "# Warning: database %s does not exist. Skipping." % (db) for table in tables: table_list.append(db + "." + table[0]) # Fail if no tables to check if not table_list: raise UtilError("No tables to check.") if verbosity > 1: print "# Checking indexes..." # Check indexes for each table in the list for table_name in table_list: tbl_options = { 'verbose': verbosity >= 1, 'get_cols': False, 'quiet': verbosity is None or verbosity < 1 } tbl = Table(source, table_name, tbl_options) exists = tbl.exists() if not exists and not skip: raise UtilError("Table %s does not exist. Use --skip " "to skip missing tables." % table_name) if exists: if not tbl.get_indexes(): if verbosity > 1: print "# Table %s is not indexed." % (table_name) else: if show_indexes: tbl.print_indexes(index_format) # Show if table has primary key if not tbl.has_primary_key(): if verbosity > 1: print "# Table %s does not contain a PRIMARY key." tbl.check_indexes(show_drops) # Show best and/or worst indexes if stats: if first_indexes is not None: tbl.show_special_indexes(index_format, first_indexes, True) if last_indexes is not None: tbl.show_special_indexes(index_format, last_indexes) if verbosity > 1: print "#" if verbosity > 1: print "# ...done."
def object_diff(server1_val, server2_val, object1, object2, options, object_type=None): """diff the definition of two objects Find the difference among two object definitions. server1_val[in] a dictionary containing connection information for the first server including: (user, password, host, port, socket) server2_val[in] a dictionary containing connection information for the second server including: (user, password, host, port, socket) object1[in] the first object in the compare in the form: (db.name) object2[in] the second object in the compare in the form: (db.name) options[in] a dictionary containing the options for the operation: (quiet, verbosity, difftype) object_type[in] type of the objects to be compared (e.g., TABLE, PROCEDURE, etc.). By default None (not defined). Returns None = objects are the same, diff[] = tables differ """ objectype = options.get("objectype", 'ALL').upper() if not object_type and objectype != 'ALL': object_type = objectype if object_type and objectype != 'ALL' and object_type != objectype: print('The object type {} is skip'.format(object_type)) return None server1, server2 = server_connect(server1_val, server2_val, object1, object2, options) force = options.get("force", None) # Get the object type if unknown considering that objects of different # types can be found with the same name. result = [] if not object_type: # Get object types of object1 sql_mode = server1.select_variable("SQL_MODE") db_name, obj_name = parse_object_name(object1, sql_mode) db = Database(server1, db_name, options) obj1_types = db.get_object_type(obj_name) if not obj1_types: msg = "The object {0} does not exist.".format(object1) if not force: raise UtilDBError(msg) print("ERROR: {0}".format(msg)) return [] # Get object types of object2 sql_mode = server2.select_variable("SQL_MODE") db_name, obj_name = parse_object_name(object2, sql_mode) db = Database(server2, db_name, options) obj2_types = db.get_object_type(obj_name) if not obj2_types: msg = "The object {0} does not exist.".format(object2) if not force: raise UtilDBError(msg) print("ERROR: {0}".format(msg)) return [] # Merge types found for both objects obj_types = set(obj1_types + obj2_types) # Diff objects considering all types found for obj_type in obj_types: res = diff_objects(server1, server2, object1, object2, options, obj_type) if res: result.append(res) else: # Diff objects of known type res = diff_objects(server1, server2, object1, object2, options, object_type) if res: result.append(res) if len(result) > 0 and options.get( "difftype", None) == 'sql' and options.get( "output", None) and options.get("output", '').endswith('.sql'): with open(options.get("output"), 'a', encoding='utf8') as fp: for res in result: if isinstance(res, list): for r in res: if r and r.strip().startswith('#'): continue fp.write('{}\n'.format(r)) else: fp.write('{}\n'.format(res)) return result if len(result) > 0 else None
def __init__(self, server1, name, options=None): """Constructor server[in] A Server object name[in] Name of table in the form (db.table) options[in] options for class: verbose, quiet, get_cols, quiet If True, do not print information messages verbose print extra data during operations (optional) (default is False) get_cols If True, get the column metadata on construction (default is False) """ if options is None: options = {} self.verbose = options.get('verbose', False) self.quiet = options.get('quiet', False) self.server = server1 # Keep table identifier considering backtick quotes if is_quoted_with_backticks(name): self.q_table = name self.q_db_name, self.q_tbl_name = Database.parse_object_name(name) self.db_name = remove_backtick_quoting(self.q_db_name) self.tbl_name = remove_backtick_quoting(self.q_tbl_name) self.table = ".".join([self.db_name, self.tbl_name]) else: self.table = name self.db_name, self.tbl_name = Database.parse_object_name(name) self.q_db_name = quote_with_backticks(self.db_name) self.q_tbl_name = quote_with_backticks(self.tbl_name) self.q_table = ".".join([self.q_db_name, self.q_tbl_name]) self.obj_type = "TABLE" self.pri_idx = None # We store each type of index in a separate list to make it easier # to manipulate self.btree_indexes = [] self.hash_indexes = [] self.rtree_indexes = [] self.fulltext_indexes = [] self.text_columns = [] self.blob_columns = [] self.column_format = None self.column_names = [] self.q_column_names = [] if options.get('get_cols', False): self.get_column_metadata() self.dest_vals = None self.storage_engine = None # Get max allowed packet res = self.server.exec_query("SELECT @@session.max_allowed_packet") if res: self.max_packet_size = res[0][0] else: self.max_packet_size = _MAXPACKET_SIZE # Watch for invalid values if self.max_packet_size > _MAXPACKET_SIZE: self.max_packet_size = _MAXPACKET_SIZE self._insert = "INSERT INTO %s.%s VALUES " self.query_options = { # Used for skipping fetch of rows 'fetch': False }
def check_index(src_val, table_args, options): """Check for duplicate or redundant indexes for one or more tables This method will examine the indexes for one or more tables and identify any indexes that are potential duplicates or redundant. It prints the equivalent DROP statements if selected. src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) table_args[in] list of tables in the form 'db.table' or 'db' options[in] dictionary of options to include: show-drops : show drop statements for dupe indexes skip : skip non-existent tables verbosity : print extra information show-indexes : show all indexes for each table index-format : index format = sql, table, tab, csv worst : show worst performing indexes best : show best performing indexes report-indexes : reports tables without PK or UK Returns bool True = success, raises UtilError if error """ # Get options show_drops = options.get("show-drops", False) skip = options.get("skip", False) verbosity = options.get("verbosity", False) show_indexes = options.get("show-indexes", False) index_format = options.get("index-format", False) stats = options.get("stats", False) first_indexes = options.get("best", None) last_indexes = options.get("worst", None) report_indexes = options.get("report-indexes", False) # Try to connect to the MySQL database server. conn_options = { 'quiet': verbosity == 1, 'version': "5.0.0", } servers = connect_servers(src_val, None, conn_options) source = servers[0] db_list = [] # list of databases table_list = [] # list of all tables to process # Build a list of objects to process # 1. start with db_list if no objects present on command line # 2. process command line options. # 3. loop through database list and add all tables # 4. check indexes # Get sql_mode value set on servers sql_mode = source.select_variable("SQL_MODE") # Perform the options check here. Loop through objects presented. for obj in table_args: m_obj = parse_object_name(obj, sql_mode) # Check if a valid database/table name is specified. if m_obj[0] is None: raise UtilError( PARSE_ERR_OBJ_NAME_FORMAT.format( obj_name=obj, option="the database/table arguments")) else: db_name, obj_name = m_obj if obj_name: # Table specified table_list.append(obj) # Else we are operating on a specific database. else: # Remove backtick quotes. db_name = remove_backtick_quoting(db_name, sql_mode) \ if is_quoted_with_backticks(db_name, sql_mode) else db_name db_list.append(db_name) # Loop through database list adding tables for db in db_list: db_source = Database(source, db) db_source.init() tables = db_source.get_db_objects("TABLE") if not tables and verbosity >= 1: print("# Warning: database %s does not exist. Skipping." % (db)) for table in tables: table_list.append("{0}.{1}".format( quote_with_backticks(db, sql_mode), quote_with_backticks(table[0], sql_mode))) # Fail if no tables to check if not table_list: raise UtilError("No tables to check.") if verbosity > 1: print("# Checking indexes...") # Check indexes for each table in the list # pylint: disable=R0101 for table_name in table_list: tbl_options = { 'verbose': verbosity >= 1, 'get_cols': False, 'quiet': verbosity is None or verbosity < 1 } tbl = Table(source, table_name, tbl_options) exists = tbl.exists() if not exists and not skip: raise UtilError("Table %s does not exist. Use --skip " "to skip missing tables." % table_name) if exists: if not tbl.get_indexes(): if verbosity > 1 or report_indexes: print("# Table %s is not indexed." % (table_name)) else: if show_indexes: tbl.print_indexes(index_format, verbosity) # Show if table has primary key if verbosity > 1 or report_indexes: if not tbl.has_primary_key(): if not tbl.has_unique_key(): print("# Table {0} does not contain neither a " "PRIMARY nor UNIQUE key.".format(table_name)) else: print("# Table {0} does not contain a PRIMARY key." "".format(table_name)) tbl.check_indexes(show_drops) # Show best and/or worst indexes if stats: if first_indexes is not None: tbl.show_special_indexes(index_format, first_indexes, True) if last_indexes is not None: tbl.show_special_indexes(index_format, last_indexes) if verbosity > 1: print("#") if verbosity > 1: print("# ...done.")
def export_data(source, src_val, db_list, options): """Produce data for the tables in a database. This method retrieves the data for each table in the databases listed in the form of BULK INSERT (SQL) statements or in a tabular form to the file specified. The valid values for the format parameter are SQL, CSV, TSV, VERITCAL, or GRID. source[in] Server instance src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) options[in] a dictionary containing the options for the copy: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, file_per_tbl, and debug) Returns bool True = success, False = error """ from mysql.utilities.common.database import Database from mysql.utilities.common.table import Table format = options.get("format", "sql") no_headers = options.get("no_headers", True) column_type = options.get("display", "brief") single = options.get("single", False) skip_blobs = options.get("skip_blobs", False) quiet = options.get("quiet", False) file_per_table = options.get("file_per_tbl", False) skip_views = options.get("skip_views", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) if options.get("all", False): rows = source.get_all_databases() for row in rows: if row[0] not in db_list: db_list.append(row[0]) # Check if database exists and user permissions on source for all databases table_lock_list = [] table_list = [] for db_name in db_list: source_db = Database(source, db_name) # Make a dictionary of the options access_options = { 'skip_views': skip_views, 'skip_procs': skip_procs, 'skip_funcs': skip_funcs, 'skip_grants': skip_grants, 'skip_events': skip_events, } # Error is source database does not exist if not source_db.exists(): raise UtilDBError("Source database does not exist - %s" % db_name, -1, db_name) source_db.check_read_access(src_val["user"], src_val["host"], access_options) # Build table list tables = source_db.get_db_objects("TABLE") for table in tables: table_list.append((db_name, table[0])) old_db = "" for table in table_list: db_name = table[0] tbl_name = "%s.%s" % (db_name, table[1]) # quote database and table name with backticks q_db_name = quote_with_backticks(db_name) q_tbl_name = "%s.%s" % (q_db_name, quote_with_backticks(table[1])) if not quiet and old_db != db_name: old_db = db_name if format == "sql": print "USE %s;" % q_db_name print "# Exporting data from %s" % db_name if file_per_table: print "# Writing table data to files." tbl_options = {'verbose': False, 'get_cols': True, 'quiet': quiet} cur_table = Table(source, q_tbl_name, tbl_options) if single and format not in ("sql", "grid", "vertical"): retrieval_mode = -1 first = True else: retrieval_mode = 1 first = False message = "# Data for table %s: " % q_tbl_name # switch for writing to files if file_per_table: if format == 'sql': file_name = tbl_name + ".sql" else: file_name = tbl_name + ".%s" % format.lower() outfile = open(file_name, "w") outfile.write(message + "\n") else: outfile = None print message for data_rows in cur_table.retrieve_rows(retrieval_mode): _export_row(data_rows, cur_table, format, single, skip_blobs, first, no_headers, outfile) if first: first = False if file_per_table: outfile.close() if not quiet: print "#...done." return True
def check_index(src_val, table_args, options): """Check for duplicate or redundant indexes for one or more tables This method will examine the indexes for one or more tables and identify any indexes that are potential duplicates or redundant. It prints the equivalent DROP statements if selected. src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) table_args[in] list of tables in the form 'db.table' or 'db' options[in] dictionary of options to include: show-drops : show drop statements for dupe indexes skip : skip non-existent tables verbosity : print extra information show-indexes : show all indexes for each table index-format : index format = sql, table, tab, csv worst : show worst performing indexes best : show best performing indexes report-indexes : reports tables without PK or UK Returns bool True = success, raises UtilError if error """ # Get options show_drops = options.get("show-drops", False) skip = options.get("skip", False) verbosity = options.get("verbosity", False) show_indexes = options.get("show-indexes", False) index_format = options.get("index-format", False) stats = options.get("stats", False) first_indexes = options.get("best", None) last_indexes = options.get("worst", None) report_indexes = options.get("report-indexes", False) # Try to connect to the MySQL database server. conn_options = { 'quiet': verbosity == 1, 'version': "5.0.0", } servers = connect_servers(src_val, None, conn_options) source = servers[0] db_list = [] # list of databases table_list = [] # list of all tables to process # Build a list of objects to process # 1. start with db_list if no objects present on command line # 2. process command line options. # 3. loop through database list and add all tables # 4. check indexes obj_name_regexp = re.compile(REGEXP_QUALIFIED_OBJ_NAME) # Perform the options check here. Loop through objects presented. for obj in table_args: m_obj = obj_name_regexp.match(obj) # Check if a valid database/table name is specified. if not m_obj: raise UtilError(PARSE_ERR_OBJ_NAME_FORMAT.format( obj_name=obj, option="the database/table arguments")) else: db_name, obj_name = m_obj.groups() if obj_name: # Table specified table_list.append(obj) # Else we are operating on a specific database. else: # Remove backtick quotes. db_name = remove_backtick_quoting(db_name) \ if is_quoted_with_backticks(db_name) else db_name db_list.append(db_name) # Loop through database list adding tables for db in db_list: db_source = Database(source, db) db_source.init() tables = db_source.get_db_objects("TABLE") if not tables and verbosity >= 1: print "# Warning: database %s does not exist. Skipping." % (db) for table in tables: table_list.append("{0}.{1}".format(quote_with_backticks(db), quote_with_backticks(table[0]))) # Fail if no tables to check if not table_list: raise UtilError("No tables to check.") if verbosity > 1: print "# Checking indexes..." # Check indexes for each table in the list for table_name in table_list: tbl_options = { 'verbose': verbosity >= 1, 'get_cols': False, 'quiet': verbosity is None or verbosity < 1 } tbl = Table(source, table_name, tbl_options) exists = tbl.exists() if not exists and not skip: raise UtilError("Table %s does not exist. Use --skip " "to skip missing tables." % table_name) if exists: if not tbl.get_indexes(): if verbosity > 1 or report_indexes: print "# Table %s is not indexed." % (table_name) else: if show_indexes: tbl.print_indexes(index_format, verbosity) # Show if table has primary key if verbosity > 1 or report_indexes: if not tbl.has_primary_key(): if not tbl.has_unique_key(): print("# Table {0} does not contain neither a " "PRIMARY nor UNIQUE key.".format(table_name)) else: print("# Table {0} does not contain a PRIMARY key." "".format(table_name)) tbl.check_indexes(show_drops) # Show best and/or worst indexes if stats: if first_indexes is not None: tbl.show_special_indexes(index_format, first_indexes, True) if last_indexes is not None: tbl.show_special_indexes(index_format, last_indexes) if verbosity > 1: print "#" if verbosity > 1: print "# ...done."
def _export_data(source, server_values, db_list, output_file, options): """Export data from the specified list of databases. This private method retrieves the data for each specified databases in SQL format (e.g., INSERT statements) or in a tabular form (GRID, TAB, CSV, VERTICAL) to the specified file. This private method does not check permissions. source[in] Server instance. server_values[in] Server connection values. db_list[in] List of databases to export. output_file[in] Output file to store the export data. options[in] Dictionary containing the options for the export: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, file_per_tbl, and debug). """ frmt = options.get("format", "sql") quiet = options.get("quiet", False) file_per_table = options.get("file_per_tbl", False) # Get tables list. table_list = [] for db_name in db_list: source_db = Database(source, db_name) # Build table list. tables = source_db.get_db_objects("TABLE") for table in tables: table_list.append((db_name, table[0])) previous_db = "" export_tbl_tasks = [] for table in table_list: # Determine start for processing table from a different database. db_name = table[0] if previous_db != db_name: previous_db = db_name if not quiet: if frmt == "sql": q_db_name = quote_with_backticks(db_name) output_file.write("USE {0};\n".format(q_db_name)) output_file.write( "# Exporting data from {0}\n".format(db_name) ) if file_per_table: output_file.write("# Writing table data to files.\n") # Print sample SOURCE command warning even in quiet mode. if file_per_table and frmt == 'sql': output_file.write("# The following are sample SOURCE commands." " If needed correct the path to match files " "location.\n") # Check multiprocess table export (only on POSIX systems). if options['multiprocess'] > 1 and os.name == 'posix': # Create export task. # Note: Server connection values are passed in the task dictionary # instead of a server instance, otherwise a multiprocessing error # is issued when assigning the task to a worker. export_task = { 'srv_con': server_values, 'table': table, 'options': options, } export_tbl_tasks.append(export_task) else: # Export data from a table (no multiprocessing). _export_table_data(source, table, output_file, options) # Print SOURCE command if --file-per-table is used and format is SQL. if file_per_table and frmt == 'sql': tbl_name = ".".join(table) output_file.write( "# SOURCE {0}\n".format(_generate_tbl_filename(tbl_name, frmt)) ) # Export tables concurrently. if export_tbl_tasks: # Create process pool. workers_pool = multiprocessing.Pool( processes=options['multiprocess'] ) # Concurrently export tables. res = workers_pool.map_async(multiprocess_tbl_export_task, export_tbl_tasks) workers_pool.close() # Get list of temporary files with the exported data. tmp_files_list = res.get() workers_pool.join() # Merge resulting temp files (if generated). for tmp_filename in tmp_files_list: if tmp_filename: tmp_file = open(tmp_filename, 'r') shutil.copyfileobj(tmp_file, output_file) tmp_file.close() os.remove(tmp_filename) if not quiet: output_file.write("#...done.\n")
def export_metadata(source, src_val, db_list, options): """Produce rows to be used to recreate objects in a database. This method retrieves the objects for each database listed in the form of CREATE (SQL) statements or in a tabular form to the file specified. The valid values for the format parameter are SQL, CSV, TSV, VERTICAL, or GRID. source[in] Server instance src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) options[in] a dictionary containing the options for the copy: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, debug, exclude_names, exclude_patterns) Returns bool True = success, False = error """ from mysql.utilities.common.database import Database from mysql.utilities.common.format import format_tabular_list from mysql.utilities.common.format import format_vertical_list format = options.get("format", "sql") no_headers = options.get("no_headers", False) column_type = options.get("display", "brief") skip_create = options.get("skip_create", False) quiet = options.get("quiet", False) skip_tables = options.get("skip_tables", False) skip_views = options.get("skip_views", False) skip_triggers = options.get("skip_triggers", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) if options.get("all", False): rows = source.get_all_databases() for row in rows: db_list.append(row[0]) # Check user permissions on source for all databases for db_name in db_list: source_db = Database(source, db_name) # Make a dictionary of the options access_options = { 'skip_views': skip_views, 'skip_procs': skip_procs, 'skip_funcs': skip_funcs, 'skip_grants': skip_grants, 'skip_events': skip_events, } source_db.check_read_access(src_val["user"], src_val["host"], access_options) for db_name in db_list: # Get a Database class instance db = Database(source, db_name, options) # Error is source database does not exist if not db.exists(): raise UtilDBError("Source database does not exist - %s" % db_name, -1, db_name) if not quiet: print "# Exporting metadata from %s" % db_name # Perform the extraction if format == "sql": db.init() # quote database name with backticks q_db_name = quote_with_backticks(db_name) if not skip_create: print "DROP DATABASE IF EXISTS %s;" % q_db_name print "CREATE DATABASE %s;" % q_db_name print "USE %s;" % q_db_name for dbobj in db.get_next_object(): if dbobj[0] == "GRANT" and not skip_grants: if not quiet: print "# Grant:" if dbobj[1][3]: create_str = "GRANT %s ON %s.%s TO %s;" % \ (dbobj[1][1], q_db_name, quote_with_backticks(dbobj[1][3]), dbobj[1][0]) else: create_str = "GRANT %s ON %s.* TO %s;" % \ (dbobj[1][1], q_db_name, dbobj[1][0]) if create_str.find("%"): create_str = re.sub("%", "%%", create_str) print create_str else: if not quiet: print "# %s: %s.%s" % (dbobj[0], db_name, dbobj[1][0]) if (dbobj[0] == "PROCEDURE" and not skip_procs) or \ (dbobj[0] == "FUNCTION" and not skip_funcs) or \ (dbobj[0] == "EVENT" and not skip_events) or \ (dbobj[0] == "TRIGGER" and not skip_triggers): print "DELIMITER ||" print "%s;" % db.get_create_statement( db_name, dbobj[1][0], dbobj[0]) if (dbobj[0] == "PROCEDURE" and not skip_procs) or \ (dbobj[0] == "FUNCTION" and not skip_funcs) or \ (dbobj[0] == "EVENT" and not skip_events) or \ (dbobj[0] == "TRIGGER" and not skip_triggers): print "||" print "DELIMITER ;" else: objects = [] if not skip_tables: objects.append("TABLE") if not skip_views: objects.append("VIEW") if not skip_triggers: objects.append("TRIGGER") if not skip_procs: objects.append("PROCEDURE") if not skip_funcs: objects.append("FUNCTION") if not skip_events: objects.append("EVENT") if not skip_grants: objects.append("GRANT") for obj_type in objects: sys.stdout.write("# %sS in %s:" % (obj_type, db_name)) if format in ('grid', 'vertical'): rows = db.get_db_objects(obj_type, column_type, True) else: rows = db.get_db_objects(obj_type, column_type, True, True) if len(rows[1]) < 1: print " (none found)" else: print # Cannot use print_list here becasue we must manipulate # the behavior of format_tabular_list list_options = {} if format == "vertical": format_vertical_list(sys.stdout, rows[0], rows[1]) elif format == "tab": list_options['print_header'] = not no_headers list_options['separator'] = '\t' format_tabular_list(sys.stdout, rows[0], rows[1], list_options) elif format == "csv": list_options['print_header'] = not no_headers list_options['separator'] = ',' format_tabular_list(sys.stdout, rows[0], rows[1], list_options) else: # default to table format format_tabular_list(sys.stdout, rows[0], rows[1]) if not quiet: print "#...done." return True
def _check_tables_structure(server1, server2, object1, object2, options, diff_type): """Check if the tables have the same structure. This method compares the tables structure ignoring the order of the columns and retrieves the differences between the table options. server1[in] first server connection. server2[in] second server connection. object1 the first object in the compare in the form: (db.name). object2 the second object in the compare in the form: (db.name). options[in] a dictionary containing the options for the operation: (quiet, verbosity, difftype, width, suppress_sql). diff_type[in] difference type. Returns a tuple (bool, list, bool) - The first tuple value is a boolean that indicates if both tables have the same structure (i.e. column definitions). The second returns the table options differences. Finally, the third is a boolean indicating if the partition options are the same. """ try: m_obj1 = re.match(REGEXP_QUALIFIED_OBJ_NAME, object1) db1, name1 = m_obj1.groups() m_obj2 = re.match(REGEXP_QUALIFIED_OBJ_NAME, object2) db2, name2 = m_obj2.groups() except: raise UtilError("Invalid object name arguments for diff_objects(): " "{0}, {1}.".format(object1, object2)) # If the second part of the object qualified name is None, then the format # is not 'db_name.obj_name' for object1 and therefore must treat it as a # database name. if not name1: return None, None, None db_1 = Database(server1, db1, options) db_2 = Database(server2, db2, options) # Get tables definitions. table_1 = db_1.get_object_definition(db1, name1, 'TABLE')[0] table_2 = db_2.get_object_definition(db2, name2, 'TABLE')[0] # Check table options. table1_opts = db_1.get_table_options(db1, name1) table2_opts = db_2.get_table_options(db2, name2) diff = _get_diff(table1_opts, table2_opts, object1, object2, diff_type) # Check if both tables have the same columns definition. # Discard column order. table_1_cols = [col[1:] for col in table_1[1]] table_2_cols = [col[1:] for col in table_2[1]] same_cols_def = set(table_1_cols) == set(table_2_cols) # Check if both tables have the same partition options. # Discard partition name. table_1_part = [part[1:] for part in table_1[2]] table_2_part = [part[1:] for part in table_2[2]] same_partition_opts = set(table_1_part) == set(table_2_part) # Return tables check results. return same_cols_def, diff, same_partition_opts
def _get_transform(server1, server2, object1, object2, options): """Get the transformation SQL statements This method generates the SQL statements to transform the destination object based on direction of the compare. server1[in] first server connection server2[in] second server connection object1 the first object in the compare in the form: (db.name) object2 the second object in the compare in the form: (db.name) options[in] a dictionary containing the options for the operation: (quiet, etc.) Returns tuple - (bool - same db name?, list of transformation statements) """ from mysql.utilities.common.database import Database from mysql.utilities.common.sql_transform import SQLTransformer obj_type = None direction = options.get("changes-for", "server1") # If there is no dot, we do not have the format 'db_name.obj_name' for # object1 and therefore must treat it as a database name. if object1.find('.') == -1: obj_type = "DATABASE" # We are working with databases so db and name need to be set # to the database name to tell the get_object_definition() method # to retrieve the database information. db1 = object1 db2 = object2 name1 = object1 name2 = object2 else: try: db1, name1 = object1.split('.') db2, name2 = object2.split('.') except: raise UtilError("Invalid object name arguments for _get_transform" "(): %s, %s." % (object1, object2)) db_1 = Database(server1, db1, options) db_2 = Database(server2, db2, options) if obj_type is None: obj_type = db_1.get_object_type(name1) transform_str = [] obj1 = db_1.get_object_definition(db1, name1, obj_type) obj2 = db_2.get_object_definition(db2, name2, obj_type) # Get the transformation based on direction. transform_str = [] same_db_name = True xform = SQLTransformer(db_1, db_2, obj1[0], obj2[0], obj_type, options.get('verbosity', 0)) differences = xform.transform_definition() if differences is not None and len(differences) > 0: transform_str.extend(differences) return transform_str
def database_compare(server1_val, server2_val, db1, db2, options): """Perform a consistency check among two databases This method performs a database consistency check among two databases which ensures the databases exist, the objects match in number and type, the row counts match for all tables, and the data for each matching tables is consistent. If any errors or differences are found, the operation stops and the difference is printed. The following steps are therefore performed: 1) check to make sure the databases exist and are the same definition 2) check to make sure the same objects exist in each database 3) for each object, ensure the object definitions match among the databases 4) for each table, ensure the row counts are the same 5) for each table, ensure the data is the same By default, the operation stops on any failure of any test. The caller can override this behavior by specifying run_all_tests = True in the options dictionary. TODO: allow the user to skip object types (e.g. --skip-triggers, et. al.) server1_val[in] a dictionary containing connection information for the first server including: (user, password, host, port, socket) server2_val[in] a dictionary containing connection information for the second server including: (user, password, host, port, socket) db1[in] the first database in the compare db2[in] the second database in the compare options[in] a dictionary containing the options for the operation: (quiet, verbosity, difftype, run_all_tests) Returns bool True if all object match, False if partial match """ _check_option_defaults(options) # Connect to servers server1, server2 = server_connect(server1_val, server2_val, db1, db2, options) # Check to see if databases exist db1_conn = Database(server1, db1, options) if not db1_conn.exists(): raise UtilDBError(_ERROR_DB_MISSING.format(db1)) db2_conn = Database(server2, db2, options) if not db2_conn.exists(): raise UtilDBError(_ERROR_DB_MISSING.format(db2)) # Print a different message is server2 is not defined if not server2_val: message = "# Checking databases {0} and {1} on server1\n#" else: message = "# Checking databases {0} on server1 and {1} on server2\n#" print(message.format(db1_conn.db_name, db2_conn.db_name)) # Check for database existence and CREATE differences _check_databases(server1, server2, db1_conn.q_db_name, db2_conn.q_db_name, options) # Get common objects and report discrepancies (in_both, differs) = _check_objects(server1, server2, db1, db2, db1_conn, db2_conn, options) success = not differs reporter = _CompareDBReport(options) reporter.print_heading() # Remaining operations can occur in a loop one for each object. for item in in_both: error_list = [] debug_msgs = [] # Set the object type obj_type = item[0] q_obj1 = "{0}.{1}".format(quote_with_backticks(db1), quote_with_backticks(item[1][0])) q_obj2 = "{0}.{1}".format(quote_with_backticks(db2), quote_with_backticks(item[1][0])) reporter.report_object(obj_type, item[1][0]) # Check for differences in CREATE errors = _compare_objects(server1, server2, q_obj1, q_obj2, reporter, options, obj_type) error_list.extend(errors) # Check row counts if obj_type == 'TABLE': errors = _check_row_counts(server1, server2, q_obj1, q_obj2, reporter, options) if len(errors) != 0: error_list.extend(errors) else: reporter.report_state("-") # Check data consistency for tables if obj_type == 'TABLE': errors, debug_msgs = _check_data_consistency( server1, server2, q_obj1, q_obj2, reporter, options) if len(errors) != 0: error_list.extend(errors) else: reporter.report_state("-") if options['verbosity'] > 0: print get_create_object(server1, q_obj1, options, obj_type) get_create_object(server2, q_obj2, options, obj_type) if debug_msgs and options['verbosity'] > 2: reporter.report_errors(debug_msgs) reporter.report_errors(error_list) # Fail if errors are found if error_list: success = False return success
def copy_db(src_val, dest_val, db_list, options): """Copy a database This method will copy a database and all of its objects and data from one server (source) to another (destination). Options are available to selectively ignore each type of object. The do_drop parameter is used to permit the copy to overwrite an existing destination database (default is to not overwrite). src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) dest_val[in] a dictionary containing connection information for the destination including: (user, password, host, port, socket) options[in] a dictionary containing the options for the copy: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, verbose, do_drop, quiet, connections, debug, exclude_names, exclude_patterns) Notes: do_drop - if True, the database on the destination will be dropped if it exists (default is False) quiet - do not print any information during operation (default is False) Returns bool True = success, False = error """ verbose = options.get("verbose", False) quiet = options.get("quiet", False) do_drop = options.get("do_drop", False) skip_views = options.get("skip_views", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) skip_data = options.get("skip_data", False) skip_triggers = options.get("skip_triggers", False) skip_tables = options.get("skip_tables", False) skip_gtid = options.get("skip_gtid", False) locking = options.get("locking", "snapshot") conn_options = { 'quiet': quiet, 'version': "5.1.30", } servers = connect_servers(src_val, dest_val, conn_options) cloning = (src_val == dest_val) or dest_val is None source = servers[0] if cloning: destination = servers[0] else: destination = servers[1] # Test if SQL_MODE is 'NO_BACKSLASH_ESCAPES' in the destination server if destination.select_variable("SQL_MODE") == "NO_BACKSLASH_ESCAPES": print( "# WARNING: The SQL_MODE in the destination server is " "'NO_BACKSLASH_ESCAPES', it will be changed temporarily " "for data insertion.") src_gtid = source.supports_gtid() == 'ON' dest_gtid = destination.supports_gtid() == 'ON' if destination else False # Get list of all databases from source if --all is specified. # Ignore system databases. if options.get("all", False): # The --all option is valid only if not cloning. if not cloning: if not quiet: print "# Including all databases." rows = source.get_all_databases() for row in rows: db_list.append((row[0], None)) # Keep same name else: raise UtilError("Cannot copy all databases on the same server.") elif not skip_gtid and src_gtid: # Check to see if this is a full copy (complete backup) all_dbs = source.exec_query("SHOW DATABASES") dbs = [db[0] for db in db_list] for db in all_dbs: if db[0].upper() in [ "MYSQL", "INFORMATION_SCHEMA", "PERFORMANCE_SCHEMA" ]: continue if not db[0] in dbs: print _GTID_BACKUP_WARNING break # Do error checking and preliminary work: # - Check user permissions on source and destination for all databases # - Check to see if executing on same server but same db name (error) # - Build list of tables to lock for copying data (if no skipping data) # - Check storage engine compatibility for db_name in db_list: source_db = Database(source, db_name[0]) if destination is None: destination = source if db_name[1] is None: db = db_name[0] else: db = db_name[1] dest_db = Database(destination, db) # Make a dictionary of the options access_options = { 'skip_views': skip_views, 'skip_procs': skip_procs, 'skip_funcs': skip_funcs, 'skip_grants': skip_grants, 'skip_events': skip_events, 'skip_triggers': skip_triggers, } source_db.check_read_access(src_val["user"], src_val["host"], access_options) # Make a dictionary containing the list of objects from source db source_objects = { "views": source_db.get_db_objects("VIEW", columns="full"), "procs": source_db.get_db_objects("PROCEDURE", columns="full"), "funcs": source_db.get_db_objects("FUNCTION", columns="full"), "events": source_db.get_db_objects("EVENT", columns="full"), "triggers": source_db.get_db_objects("TRIGGER", columns="full"), } dest_db.check_write_access(dest_val['user'], dest_val['host'], access_options, source_objects, do_drop) # Error is source db and destination db are the same and we're cloning if destination == source and db_name[0] == db_name[1]: raise UtilError("Destination database name is same as " "source - source = %s, destination = %s" % (db_name[0], db_name[1])) # Error is source database does not exist if not source_db.exists(): raise UtilError("Source database does not exist - %s" % db_name[0]) # Check storage engines check_engine_options(destination, options.get("new_engine", None), options.get("def_engine", None), False, options.get("quiet", False)) # Get replication commands if rpl_mode specified. # if --rpl specified, dump replication initial commands rpl_info = None # Turn off foreign keys if they were on at the start destination.disable_foreign_key_checks(True) # Get GTID commands if not skip_gtid: gtid_info = get_gtid_commands(source) if src_gtid and not dest_gtid: print _NON_GTID_WARNING % ("destination", "source", "to") elif not src_gtid and dest_gtid: print _NON_GTID_WARNING % ("source", "destination", "from") else: gtid_info = None if src_gtid and not cloning: print _GTID_WARNING # If cloning, turn off gtid generation if gtid_info and cloning: gtid_info = None # if GTIDs enabled, write the GTID commands if gtid_info and dest_gtid: # Check GTID version for complete feature support destination.check_gtid_version() # Check the gtid_purged value too destination.check_gtid_executed() for cmd in gtid_info[0]: print "# GTID operation:", cmd destination.exec_query(cmd, {'fetch': False, 'commit': False}) if options.get("rpl_mode", None): new_opts = options.copy() new_opts['multiline'] = False new_opts['strict'] = True rpl_info = get_change_master_command(src_val, new_opts) destination.exec_query("STOP SLAVE", {'fetch': False, 'commit': False}) # Copy (create) objects. # We need to delay trigger and events to after data is loaded new_opts = options.copy() new_opts['skip_triggers'] = True new_opts['skip_events'] = True # Get the table locks unless we are cloning with lock-all if not (cloning and locking == 'lock-all'): my_lock = get_copy_lock(source, db_list, options, True) _copy_objects(source, destination, db_list, new_opts) # If we are cloning, take the write locks prior to copying data if cloning and locking == 'lock-all': my_lock = get_copy_lock(source, db_list, options, True, cloning) # Copy tables data if not skip_data and not skip_tables: # Copy tables for db_name in db_list: # Get a Database class instance db = Database(source, db_name[0], options) # Perform the copy # Note: No longer use threads, use multiprocessing instead. db.init() db.copy_data(db_name[1], options, destination, connections=1, src_con_val=src_val, dest_con_val=dest_val) # if cloning with lock-all unlock here to avoid system table lock conflicts if cloning and locking == 'lock-all': my_lock.unlock() # Create triggers for all databases if not skip_triggers: new_opts = options.copy() new_opts['skip_tables'] = True new_opts['skip_views'] = True new_opts['skip_procs'] = True new_opts['skip_funcs'] = True new_opts['skip_events'] = True new_opts['skip_grants'] = True new_opts['skip_create'] = True _copy_objects(source, destination, db_list, new_opts, False, False) # Create events for all databases if not skip_events: new_opts = options.copy() new_opts['skip_tables'] = True new_opts['skip_views'] = True new_opts['skip_procs'] = True new_opts['skip_funcs'] = True new_opts['skip_triggers'] = True new_opts['skip_grants'] = True new_opts['skip_create'] = True _copy_objects(source, destination, db_list, new_opts, False, False) if not (cloning and locking == 'lock-all'): my_lock.unlock() # if GTIDs enabled, write the GTID-related commands if gtid_info and dest_gtid: print "# GTID operation:", gtid_info[1] destination.exec_query(gtid_info[1]) if options.get("rpl_mode", None): for cmd in rpl_info[_RPL_COMMANDS]: if cmd[0] == '#' and not quiet: print cmd else: if verbose: print cmd destination.exec_query(cmd) destination.exec_query("START SLAVE;") # Turn on foreign keys if they were on at the start destination.disable_foreign_key_checks(False) if not quiet: print "#...done." return True
def copy_db(src_val, dest_val, db_list, options): """Copy a database This method will copy a database and all of its objects and data from one server (source) to another (destination). Options are available to selectively ignore each type of object. The force parameter is used to permit the copy to overwrite an existing destination database (default is to not overwrite). src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) dest_val[in] a dictionary containing connection information for the destination including: (user, password, host, port, socket) options[in] a dictionary containing the options for the copy: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, verbose, force, quiet, connections, debug, exclude_names, exclude_patterns) Notes: force - if True, the database on the destination will be dropped if it exists (default is False) quiet - do not print any information during operation (default is False) Returns bool True = success, False = error """ verbose = options.get("verbose", False) quiet = options.get("quiet", False) skip_views = options.get("skip_views", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) skip_data = options.get("skip_data", False) skip_triggers = options.get("skip_triggers", False) skip_tables = options.get("skip_tables", False) skip_gtid = options.get("skip_gtid", False) locking = options.get("locking", "snapshot") conn_options = { 'quiet': quiet, 'version': "5.1.30", } servers = connect_servers(src_val, dest_val, conn_options) cloning = (src_val == dest_val) or dest_val is None source = servers[0] if cloning: destination = servers[0] else: destination = servers[1] src_gtid = source.supports_gtid() == 'ON' dest_gtid = destination.supports_gtid() == 'ON'if destination else False # Get list of all databases from source if --all is specified. # Ignore system databases. if options.get("all", False): # The --all option is valid only if not cloning. if not cloning: if not quiet: print "# Including all databases." rows = source.get_all_databases() for row in rows: db_list.append((row[0], None)) # Keep same name else: raise UtilError("Cannot copy all databases on the same server.") elif not skip_gtid and src_gtid: # Check to see if this is a full copy (complete backup) all_dbs = source.exec_query("SHOW DATABASES") dbs = [db[0] for db in db_list] for db in all_dbs: if db[0].upper() in ["MYSQL", "INFORMATION_SCHEMA", "PERFORMANCE_SCHEMA"]: continue if not db[0] in dbs: print _GTID_BACKUP_WARNING break # Do error checking and preliminary work: # - Check user permissions on source and destination for all databases # - Check to see if executing on same server but same db name (error) # - Build list of tables to lock for copying data (if no skipping data) # - Check storage engine compatibility for db_name in db_list: source_db = Database(source, db_name[0]) if destination is None: destination = source if db_name[1] is None: db = db_name[0] else: db = db_name[1] dest_db = Database(destination, db) # Make a dictionary of the options access_options = { 'skip_views': skip_views, 'skip_procs': skip_procs, 'skip_funcs': skip_funcs, 'skip_grants': skip_grants, 'skip_events': skip_events, } source_db.check_read_access(src_val["user"], src_val["host"], access_options) dest_db.check_write_access(dest_val['user'], dest_val['host'], access_options) # Error is source db and destination db are the same and we're cloning if destination == source and db_name[0] == db_name[1]: raise UtilError("Destination database name is same as " "source - source = %s, destination = %s" % (db_name[0], db_name[1])) # Error is source database does not exist if not source_db.exists(): raise UtilError("Source database does not exist - %s" % db_name[0]) # Check storage engines check_engine_options(destination, options.get("new_engine", None), options.get("def_engine", None), False, options.get("quiet", False)) # Get replication commands if rpl_mode specified. # if --rpl specified, dump replication initial commands rpl_info = None # Turn off foreign keys if they were on at the start destination.disable_foreign_key_checks(True) # Get GTID commands if not skip_gtid: gtid_info = get_gtid_commands(source) if src_gtid and not dest_gtid: print _NON_GTID_WARNING % ("destination", "source", "to") elif not src_gtid and dest_gtid: print _NON_GTID_WARNING % ("source", "destination", "from") else: gtid_info = None if src_gtid and not cloning: print _GTID_WARNING # If cloning, turn off gtid generation if gtid_info and cloning: gtid_info = None # if GTIDs enabled, write the GTID commands if gtid_info and dest_gtid: # Check GTID version for complete feature support destination.check_gtid_version() # Check the gtid_purged value too destination.check_gtid_executed() for cmd in gtid_info[0]: print "# GTID operation:", cmd destination.exec_query(cmd, {'fetch': False, 'commit': False}) if options.get("rpl_mode", None): new_opts = options.copy() new_opts['multiline'] = False new_opts['strict'] = True rpl_info = get_change_master_command(src_val, new_opts) destination.exec_query("STOP SLAVE", {'fetch': False, 'commit': False}) # Copy (create) objects. # We need to delay trigger and events to after data is loaded new_opts = options.copy() new_opts['skip_triggers'] = True new_opts['skip_events'] = True # Get the table locks unless we are cloning with lock-all if not (cloning and locking == 'lock-all'): my_lock = get_copy_lock(source, db_list, options, True) _copy_objects(source, destination, db_list, new_opts) # If we are cloning, take the write locks prior to copying data if cloning and locking == 'lock-all': my_lock = get_copy_lock(source, db_list, options, True, cloning) # Copy tables data if not skip_data and not skip_tables: # Copy tables for db_name in db_list: # Get a Database class instance db = Database(source, db_name[0], options) # Perform the copy # Note: No longer use threads, use multiprocessing instead. db.init() db.copy_data(db_name[1], options, destination, connections=1, src_con_val=src_val, dest_con_val=dest_val) # if cloning with lock-all unlock here to avoid system table lock conflicts if cloning and locking == 'lock-all': my_lock.unlock() # Create triggers for all databases if not skip_triggers: new_opts = options.copy() new_opts['skip_tables'] = True new_opts['skip_views'] = True new_opts['skip_procs'] = True new_opts['skip_funcs'] = True new_opts['skip_events'] = True new_opts['skip_grants'] = True new_opts['skip_create'] = True _copy_objects(source, destination, db_list, new_opts, False, False) # Create events for all databases if not skip_events: new_opts = options.copy() new_opts['skip_tables'] = True new_opts['skip_views'] = True new_opts['skip_procs'] = True new_opts['skip_funcs'] = True new_opts['skip_triggers'] = True new_opts['skip_grants'] = True new_opts['skip_create'] = True _copy_objects(source, destination, db_list, new_opts, False, False) if not (cloning and locking == 'lock-all'): my_lock.unlock() # if GTIDs enabled, write the GTID-related commands if gtid_info and dest_gtid: print "# GTID operation:", gtid_info[1] destination.exec_query(gtid_info[1]) if options.get("rpl_mode", None): for cmd in rpl_info[_RPL_COMMANDS]: if cmd[0] == '#' and not quiet: print cmd else: if verbose: print cmd destination.exec_query(cmd) destination.exec_query("START SLAVE;") # Turn on foreign keys if they were on at the start destination.disable_foreign_key_checks(False) if not quiet: print "#...done." return True
def object_diff(server1_val, server2_val, object1, object2, options, object_type=None): """diff the definition of two objects Find the difference among two object definitions. server1_val[in] a dictionary containing connection information for the first server including: (user, password, host, port, socket) server2_val[in] a dictionary containing connection information for the second server including: (user, password, host, port, socket) object1[in] the first object in the compare in the form: (db.name) object2[in] the second object in the compare in the form: (db.name) options[in] a dictionary containing the options for the operation: (quiet, verbosity, difftype) object_type[in] type of the objects to be compared (e.g., TABLE, PROCEDURE, etc.). By default None (not defined). Returns None = objects are the same, diff[] = tables differ """ server1, server2 = server_connect(server1_val, server2_val, object1, object2, options) force = options.get("force", None) # Get the object type if unknown considering that objects of different # types can be found with the same name. if not object_type: # Get object types of object1 sql_mode = server1.select_variable("SQL_MODE") db_name, obj_name = parse_object_name(object1, sql_mode) db = Database(server1, db_name, options) obj1_types = db.get_object_type(obj_name) if not obj1_types: msg = "The object {0} does not exist.".format(object1) if not force: raise UtilDBError(msg) print("ERROR: {0}".format(msg)) return [] # Get object types of object2 sql_mode = server2.select_variable("SQL_MODE") db_name, obj_name = parse_object_name(object2, sql_mode) db = Database(server2, db_name, options) obj2_types = db.get_object_type(obj_name) if not obj2_types: msg = "The object {0} does not exist.".format(object2) if not force: raise UtilDBError(msg) print("ERROR: {0}".format(msg)) return [] # Merge types found for both objects obj_types = set(obj1_types + obj2_types) # Diff objects considering all types found result = [] for obj_type in obj_types: res = diff_objects(server1, server2, object1, object2, options, obj_type) if res: result.append(res) return result if len(result) > 0 else None else: # Diff objects of known type return diff_objects(server1, server2, object1, object2, options, object_type)
def _export_metadata(source, db_list, output_file, options): """Export metadata from the specified list of databases. This private method retrieves the objects metadata for each database listed in the form of CREATE (SQL) statements or in a tabular form (GRID, TAB, CSV, VERTICAL) to the specified file. This private method does not check permissions. source[in] Server instance. db_list[in] List of databases to export. output_file[in] Output file to store the metadata information. options[in] Dictionary containing the options for the export: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, debug, exclude_names, exclude_patterns) """ frmt = options.get("format", "sql") no_headers = options.get("no_headers", False) column_type = options.get("display", "brief") quiet = options.get("quiet", False) skip_create = options.get("skip_create", False) skip_tables = options.get("skip_tables", False) skip_views = options.get("skip_views", False) skip_triggers = options.get("skip_triggers", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) sql_mode = source.select_variable("SQL_MODE") for db_name in db_list: # Get a Database class instance db = Database(source, db_name, options) # Export database metadata if not quiet: output_file.write("# Exporting metadata from {0}\n".format( db.q_db_name)) # Perform the extraction if frmt == "sql": db.init() if not skip_create: output_file.write("DROP DATABASE IF EXISTS {0};\n".format( db.q_db_name)) output_file.write("CREATE DATABASE {0};\n".format( db.q_db_name)) output_file.write("USE {0};\n".format(db.q_db_name)) for dbobj in db.get_next_object(): if dbobj[0] == "GRANT" and not skip_grants: if not quiet: output_file.write("# Grant:\n") if dbobj[1][3]: create_str = "GRANT {0} ON {1}.{2} TO {3};\n".format( dbobj[1][1], db.q_db_name, quote_with_backticks(dbobj[1][3], sql_mode), dbobj[1][0]) else: create_str = "GRANT {0} ON {1}.* TO {2};\n".format( dbobj[1][1], db.q_db_name, dbobj[1][0]) output_file.write(create_str) else: if not quiet: output_file.write("# {0}: {1}.{2}\n".format( dbobj[0], db.q_db_name, quote_with_backticks(dbobj[1][0], sql_mode))) if (dbobj[0] == "PROCEDURE" and not skip_procs) or \ (dbobj[0] == "FUNCTION" and not skip_funcs) or \ (dbobj[0] == "EVENT" and not skip_events) or \ (dbobj[0] == "TRIGGER" and not skip_triggers): output_file.write("DELIMITER ||\n") output_file.write("{0};\n".format( db.get_create_statement(db.db_name, dbobj[1][0], dbobj[0]))) if (dbobj[0] == "PROCEDURE" and not skip_procs) or \ (dbobj[0] == "FUNCTION" and not skip_funcs) or \ (dbobj[0] == "EVENT" and not skip_events) or \ (dbobj[0] == "TRIGGER" and not skip_triggers): output_file.write("||\n") output_file.write("DELIMITER ;\n") else: objects = [] if not skip_tables: objects.append("TABLE") if not skip_funcs: objects.append("FUNCTION") if not skip_procs: objects.append("PROCEDURE") if not skip_views: objects.append("VIEW") if not skip_triggers: objects.append("TRIGGER") if not skip_events: objects.append("EVENT") if not skip_grants: objects.append("GRANT") for obj_type in objects: output_file.write("# {0}S in {1}:".format( obj_type, db.q_db_name)) if frmt in ('grid', 'vertical'): rows = db.get_db_objects(obj_type, column_type, True) else: rows = db.get_db_objects(obj_type, column_type, True, True) if len(rows[1]) < 1: output_file.write(" (none found)\n") else: output_file.write("\n") # Cannot use print_list here because we must manipulate # the behavior of format_tabular_list. list_options = {} if frmt == "vertical": format_vertical_list(output_file, rows[0], rows[1]) elif frmt == "tab": list_options['print_header'] = not no_headers list_options['separator'] = '\t' format_tabular_list(output_file, rows[0], rows[1], list_options) elif frmt == "csv": list_options['print_header'] = not no_headers list_options['separator'] = ',' format_tabular_list(output_file, rows[0], rows[1], list_options) else: # default to table format format_tabular_list(output_file, rows[0], rows[1]) if not quiet: output_file.write("#...done.\n")
def database_compare(server1_val, server2_val, db1, db2, options): """Perform a consistency check among two databases This method performs a database consistency check among two databases which ensures the databases exist, the objects match in number and type, the row counts match for all tables, and the data for each matching tables is consistent. If any errors or differences are found, the operation stops and the difference is printed. The following steps are therefore performed: 1) check to make sure the databases exist and are the same definition 2) check to make sure the same objects exist in each database 3) for each object, ensure the object definitions match among the databases 4) for each table, ensure the row counts are the same 5) for each table, ensure the data is the same By default, the operation stops on any failure of any test. The caller can override this behavior by specifying run_all_tests = True in the options dictionary. TODO: allow the user to skip object types (e.g. --skip-triggers, et. al.) server1_val[in] a dictionary containing connection information for the first server including: (user, password, host, port, socket) server2_val[in] a dictionary containing connection information for the second server including: (user, password, host, port, socket) db1[in] the first database in the compare db2[in] the second database in the compare options[in] a dictionary containing the options for the operation: (quiet, verbosity, difftype, run_all_tests) Returns bool True if all object match, False if partial match """ _check_option_defaults(options) # Connect to servers server1, server2 = server_connect(server1_val, server2_val, db1, db2, options) # Check to see if databases exist db1_conn = Database(server1, db1, options) if not db1_conn.exists(): raise UtilDBError(_ERROR_DB_MISSING.format(db1)) db2_conn = Database(server2, db2, options) if not db2_conn.exists(): raise UtilDBError(_ERROR_DB_MISSING.format(db2)) # Print a different message is server2 is not defined if not server2_val: message = "# Checking databases {0} and {1} on server1\n#" else: message = "# Checking databases {0} on server1 and {1} on server2\n#" print(message.format(db1_conn.db_name, db2_conn.db_name)) # Check for database existence and CREATE differences _check_databases(server1, server2, db1_conn.q_db_name, db2_conn.q_db_name, options) # Get common objects and report discrepancies (in_both, differs) = _check_objects(server1, server2, db1, db2, db1_conn, db2_conn, options) success = not differs reporter = _CompareDBReport(options) reporter.print_heading() # Remaining operations can occur in a loop one for each object. for item in in_both: error_list = [] # Set the object type obj_type = item[0] obj1 = "{0}.{1}".format(db1, item[1][0]) obj2 = "{0}.{1}".format(db2, item[1][0]) q_obj1 = "{0}.{1}".format(quote_with_backticks(db1), quote_with_backticks(item[1][0])) q_obj2 = "{0}.{1}".format(quote_with_backticks(db2), quote_with_backticks(item[1][0])) reporter.report_object(obj_type, item[1][0]) # Check for differences in CREATE errors = _compare_objects(server1, server2, q_obj1, q_obj2, reporter, options, obj_type) error_list.extend(errors) # Check row counts if obj_type == 'TABLE': errors = _check_row_counts(server1, server2, q_obj1, q_obj2, reporter, options) if len(errors) != 0: error_list.extend(errors) else: reporter.report_state("-") # Check data consistency for tables if obj_type == 'TABLE': errors = _check_data_consistency(server1, server2, q_obj1, q_obj2, reporter, options) if len(errors) != 0: error_list.extend(errors) else: reporter.report_state("-") if options['verbosity'] > 0: print get_create_object(server1, obj1, options, obj_type) get_create_object(server2, obj2, options, obj_type) reporter.report_errors(error_list) # Fail if errors are found if error_list: success = False return success
def get_copy_lock(server, db_list, options, include_mysql=False, cloning=False): """Get an instance of the Lock class with a standard copy (read) lock This method creates an instance of the Lock class using the lock type specified in the options. It is used to initiate the locks for the copy and related operations. server[in] Server instance for locking calls db_list[in] list of database names options[in] option dictionary Must include the skip_* options for copy and export include_mysql[in] if True, include the mysql tables for copy operation cloning[in] if True, create lock tables with WRITE on dest db Default = False Returns Lock - Lock class instance """ rpl_mode = options.get("rpl_mode", None) locking = options.get('locking', 'snapshot') # Determine if we need to use FTWRL. There are two conditions: # - running on master (rpl_mode = 'master') # - using locking = 'lock-all' and rpl_mode present if (rpl_mode in ["master", "both"]) or \ (rpl_mode and locking == 'lock-all'): new_opts = options.copy() new_opts['locking'] = 'flush' lock = Lock(server, [], new_opts) # if this is a lock-all type and not replication operation, # find all tables and lock them # pylint: disable=R0101 elif locking == 'lock-all': table_lock_list = [] # Build table lock list for db_name in db_list: db = db_name[0] if isinstance(db_name, tuple) else db_name source_db = Database(server, db) tables = source_db.get_db_objects("TABLE") for table in tables: table_lock_list.append(("{0}.{1}".format(db, table[0]), 'READ')) # Cloning requires issuing WRITE locks because we use same # conn. # Non-cloning will issue WRITE lock on a new destination conn. if cloning: if db_name[1] is None: db_clone = db_name[0] else: db_clone = db_name[1] # For cloning, we use the same connection so we need to # lock the destination tables with WRITE. table_lock_list.append( ("{0}.{1}".format(db_clone, table[0]), 'WRITE')) # We must include views for server version 5.5.3 and higher if server.check_version_compat(5, 5, 3): tables = source_db.get_db_objects("VIEW") for table in tables: table_lock_list.append( ("{0}.{1}".format(db, table[0]), 'READ')) # Cloning requires issuing WRITE locks because we use same # conn. # Non-cloning will issue WRITE lock on a new destination # conn. if cloning: if db_name[1] is None: db_clone = db_name[0] else: db_clone = db_name[1] # For cloning, we use the same connection so we need to # lock the destination tables with WRITE. table_lock_list.append( ("{0}.{1}".format(db_clone, table[0]), 'WRITE')) # Now add mysql tables if include_mysql: # Don't lock proc tables if no procs of funcs are being read if not options.get('skip_procs', False) and \ not options.get('skip_funcs', False): table_lock_list.append(("mysql.proc", 'READ')) table_lock_list.append(("mysql.procs_priv", 'READ')) # Don't lock event table if events are skipped if not options.get('skip_events', False): table_lock_list.append(("mysql.event", 'READ')) lock = Lock(server, table_lock_list, options) # Use default or no locking option else: lock = Lock(server, [], options) return lock
def export_data(source, src_val, db_list, options): """Produce data for the tables in a database. This method retrieves the data for each table in the databases listed in the form of BULK INSERT (SQL) statements or in a tabular form to the file specified. The valid values for the format parameter are SQL, CSV, TSV, VERITCAL, or GRID. source[in] Server instance src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) options[in] a dictionary containing the options for the copy: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, file_per_tbl, and debug) Returns bool True = success, False = error """ from mysql.utilities.common.database import Database from mysql.utilities.common.table import Table format = options.get("format", "sql") no_headers = options.get("no_headers", True) column_type = options.get("display", "brief") single = options.get("single", False) skip_blobs = options.get("skip_blobs", False) quiet = options.get("quiet", False) file_per_table = options.get("file_per_tbl", False) skip_views = options.get("skip_views", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) if options.get("all", False): rows = source.get_all_databases() for row in rows: if row[0] not in db_list: db_list.append(row[0]) # Check if database exists and user permissions on source for all databases table_lock_list = [] table_list = [] for db_name in db_list: source_db = Database(source, db_name) # Make a dictionary of the options access_options = { 'skip_views' : skip_views, 'skip_procs' : skip_procs, 'skip_funcs' : skip_funcs, 'skip_grants' : skip_grants, 'skip_events' : skip_events, } # Error is source database does not exist if not source_db.exists(): raise UtilDBError("Source database does not exist - %s" % db_name, -1, db_name) source_db.check_read_access(src_val["user"], src_val["host"], access_options) # Build table list tables = source_db.get_db_objects("TABLE") for table in tables: table_list.append((db_name, table[0])) old_db = "" for table in table_list: db_name = table[0] tbl_name = "%s.%s" % (db_name, table[1]) # quote database and table name with backticks q_db_name = quote_with_backticks(db_name) q_tbl_name = "%s.%s" % (q_db_name, quote_with_backticks(table[1])) if not quiet and old_db != db_name: old_db = db_name if format == "sql": print "USE %s;" % q_db_name print "# Exporting data from %s" % db_name if file_per_table: print "# Writing table data to files." tbl_options = { 'verbose' : False, 'get_cols' : True, 'quiet' : quiet } cur_table = Table(source, q_tbl_name, tbl_options) if single and format not in ("sql", "grid", "vertical"): retrieval_mode = -1 first = True else: retrieval_mode = 1 first = False message = "# Data for table %s: " % q_tbl_name # switch for writing to files if file_per_table: if format == 'sql': file_name = tbl_name + ".sql" else: file_name = tbl_name + ".%s" % format.lower() outfile = open(file_name, "w") outfile.write(message + "\n") else: outfile = None print message for data_rows in cur_table.retrieve_rows(retrieval_mode): _export_row(data_rows, cur_table, format, single, skip_blobs, first, no_headers, outfile) if first: first = False if file_per_table: outfile.close() if not quiet: print "#...done." return True
def validate_obj_type_dict(server, obj_type_dict): """Validates the dictionary of objects against the specified server This function builds a dict with the types of the objects in obj_type_dict, filtering out non existing databases and objects. Returns a dictionary with only the existing objects, using object_types as keys and as values a list of tuples (<DB NAME>, <OBJ_NAME>). """ valid_obj_dict = defaultdict(list) server_dbs = set(row[0] for row in server.get_all_databases( ignore_internal_dbs=False)) argument_dbs = set(obj_type_dict.keys()) # Get non existing_databases and dbs to check non_existing_dbs = argument_dbs.difference(server_dbs) dbs_to_check = server_dbs.intersection(argument_dbs) if non_existing_dbs: if len(non_existing_dbs) > 1: plurals = ('s', '', 'them') else: plurals = ('', 'es', 'it') print('# WARNING: specified database{0} do{1} not ' 'exist on base server and will be skipped along ' 'any tables and routines belonging to {2}: ' '{3}.'.format(plurals[0], plurals[1], plurals[2], ", ".join(non_existing_dbs))) # Now for each db that actually exists, get the type of the specified # objects for db_name in dbs_to_check: db = Database(server, db_name) # quote database name if necessary quoted_db_name = db_name if not is_quoted_with_backticks(db_name): quoted_db_name = quote_with_backticks(db_name) for obj_name in obj_type_dict[db_name]: if obj_name is None: # We must consider the database itself valid_obj_dict[DATABASE_TYPE].append((quoted_db_name, quoted_db_name)) else: # get quoted name for obj_name quoted_obj_name = obj_name if not is_quoted_with_backticks(obj_name): quoted_obj_name = quote_with_backticks(obj_name) # Test if the object exists and if it does, test if it # is one of the supported object types, else # print a warning and skip the object obj_type = db.get_object_type(obj_name) if obj_type is None: print("# WARNING: specified object does not exist. " "{0}.{1} will be skipped." "".format(quoted_db_name, quoted_obj_name)) elif 'PROCEDURE' in obj_type or 'FUNCTION' in obj_type: valid_obj_dict[ROUTINE_TYPE].append((quoted_db_name, quoted_obj_name)) elif 'TABLE' in obj_type: valid_obj_dict[TABLE_TYPE].append((quoted_db_name, quoted_obj_name)) else: print('# WARNING: specified object is not supported ' '(not a DATABASE, FUNCTION, PROCEDURE or TABLE),' ' as such it will be skipped: {0}.{1}.' ''.format(quoted_db_name, quoted_obj_name)) return valid_obj_dict
def copy_db(src_val, dest_val, db_list, options): """Copy a database This method will copy a database and all of its objects and data from one server (source) to another (destination). Options are available to selectively ignore each type of object. The force parameter is used to permit the copy to overwrite an existing destination database (default is to not overwrite). src_val[in] a dictionary containing connection information for the source including: (user, password, host, port, socket) dest_val[in] a dictionary containing connection information for the destination including: (user, password, host, port, socket) options[in] a dictionary containing the options for the copy: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, verbose, force, quiet, connections, debug, exclude_names, exclude_patterns) Notes: force - if True, the database on the destination will be dropped if it exists (default is False) quiet - do not print any information during operation (default is False) Returns bool True = success, False = error """ from mysql.utilities.common.database import Database from mysql.utilities.common.options import check_engine_options from mysql.utilities.common.server import connect_servers from mysql.utilities.command.dbexport import get_change_master_command verbose = options.get("verbose", False) quiet = options.get("quiet", False) skip_views = options.get("skip_views", False) skip_procs = options.get("skip_procs", False) skip_funcs = options.get("skip_funcs", False) skip_events = options.get("skip_events", False) skip_grants = options.get("skip_grants", False) skip_data = options.get("skip_data", False) skip_triggers = options.get("skip_triggers", False) skip_tables = options.get("skip_tables", False) locking = options.get("locking", "snapshot") rpl_info = ([], None) conn_options = { 'quiet' : quiet, 'version' : "5.1.30", } servers = connect_servers(src_val, dest_val, conn_options) source = servers[0] destination = servers[1] cloning = (src_val == dest_val) or dest_val is None # Get list of all databases from source if --all is specified. # Ignore system databases. if options.get("all", False): # The --all option is valid only if not cloning. if not cloning: if not quiet: print "# Including all databases." rows = source.get_all_databases() for row in rows: db_list.append((row[0], None)) # Keep same name else: raise UtilError("Cannot copy all databases on the same server.") # Do error checking and preliminary work: # - Check user permissions on source and destination for all databases # - Check to see if executing on same server but same db name (error) # - Build list of tables to lock for copying data (if no skipping data) # - Check storage engine compatibility for db_name in db_list: source_db = Database(source, db_name[0]) if destination is None: destination = source if db_name[1] is None: db = db_name[0] else: db = db_name[1] dest_db = Database(destination, db) # Make a dictionary of the options access_options = { 'skip_views' : skip_views, 'skip_procs' : skip_procs, 'skip_funcs' : skip_funcs, 'skip_grants' : skip_grants, 'skip_events' : skip_events, } source_db.check_read_access(src_val["user"], src_val["host"], access_options) dest_db.check_write_access(dest_val['user'], dest_val['host'], access_options) # Error is source db and destination db are the same and we're cloning if destination == source and db_name[0] == db_name[1]: raise UtilError("Destination database name is same as " "source - source = %s, destination = %s" % (db_name[0], db_name[1])) # Error is source database does not exist if not source_db.exists(): raise UtilError("Source database does not exist - %s" % db_name[0]) # Check storage engines check_engine_options(destination, options.get("new_engine", None), options.get("def_engine", None), False, options.get("quiet", False)) # Get replication commands if rpl_mode specified. # if --rpl specified, dump replication initial commands if options.get("rpl_mode", None): new_opts = options.copy() new_opts['multiline'] = False new_opts['strict'] = True rpl_info = get_change_master_command(src_val, new_opts) destination.exec_query("STOP SLAVE;") # Copy objects # We need to delay trigger and events to after data is loaded new_opts = options.copy() new_opts['skip_triggers'] = True new_opts['skip_events'] = True # Get the table locks unless we are cloning with lock-all if not (cloning and locking == 'lock-all'): my_lock = get_copy_lock(source, db_list, options, True) _copy_objects(source, destination, db_list, new_opts) # If we are cloning, take the write locks prior to copying data if cloning and locking == 'lock-all': my_lock = get_copy_lock(source, db_list, options, True, cloning) # Copy data if not skip_data and not skip_tables: # Copy tables for db_name in db_list: # Get a Database class instance db = Database(source, db_name[0], options) # Perform the copy db.init() db.copy_data(db_name[1], options, destination, options.get("threads", False)) # if cloning with lock-all unlock here to avoid system table lock conflicts if cloning and locking == 'lock-all': my_lock.unlock() # Create triggers for all databases if not skip_triggers: new_opts = options.copy() new_opts['skip_tables'] = True new_opts['skip_views'] = True new_opts['skip_procs'] = True new_opts['skip_funcs'] = True new_opts['skip_events'] = True new_opts['skip_grants'] = True new_opts['skip_create'] = True _copy_objects(source, destination, db_list, new_opts, False, False) # Create events for all databases if not skip_events: new_opts = options.copy() new_opts['skip_tables'] = True new_opts['skip_views'] = True new_opts['skip_procs'] = True new_opts['skip_funcs'] = True new_opts['skip_triggers'] = True new_opts['skip_grants'] = True new_opts['skip_create'] = True _copy_objects(source, destination, db_list, new_opts, False, False) if not (cloning and locking == 'lock-all'): my_lock.unlock() if options.get("rpl_mode", None): for cmd in rpl_info[_RPL_COMMANDS]: if cmd[0] == '#' and not quiet: print cmd else: if verbose: print cmd destination.exec_query(cmd) destination.exec_query("START SLAVE;") if not quiet: print "#...done." return True
def import_file(dest_val, file_name, options): """Import a file This method reads a file and, if needed, transforms the file into discrete SQL statements for execution on the destination server. It accepts any of the formal structured files produced by the mysqlexport utility including formats SQL, CSV, TAB, GRID, and VERTICAL. It will read these files and skip or include the definitions or data as specified in the options. An error is raised for any conversion errors or errors while executing the statements. Users are highly encouraged to use the --dryrun option which will print the SQL statements without executing them. dest_val[in] a dictionary containing connection information for the destination including: (user, password, host, port, socket) file_name[in] name (and path) of the file to import options[in] a dictionary containing the options for the import: (skip_tables, skip_views, skip_triggers, skip_procs, skip_funcs, skip_events, skip_grants, skip_create, skip_data, no_header, display, format, and debug) Returns bool True = success, False = error """ from mysql.utilities.common.database import Database from mysql.utilities.common.options import check_engine_options from mysql.utilities.common.table import Table from mysql.utilities.common.server import connect_servers # Helper method to dig through the definitions for create statements def _process_definitions(statements, table_col_list, db_name): # First, get the SQL strings sql_strs = _build_create_objects(obj_type, db_name, definitions) statements.extend(sql_strs) # Now, save the column list col_list = _build_col_metadata(obj_type, definitions) if len(col_list) > 0: table_col_list.extend(col_list) def _process_data(tbl_name, statements, columns, table_col_list, table_rows, skip_blobs): # if there is data here, build bulk inserts # First, create table reference, then call insert_rows() tbl = Table(destination, tbl_name) # Need to check to see if table exists! if tbl.exists(): tbl.get_column_metadata() col_meta = True elif len(table_col_list) > 0: col_meta = _get_column_metadata(tbl, table_col_list) else: fix_cols = [] fix_cols.append((tbl.tbl_name, columns)) col_meta = _get_column_metadata(tbl, fix_cols) if not col_meta: raise UtilError("Cannot build bulk insert statements without " "the table definition.") ins_strs = tbl.make_bulk_insert(table_rows, tbl.q_db_name) if len(ins_strs[0]) > 0: statements.extend(ins_strs[0]) if len(ins_strs[1]) > 0 and not skip_blobs: for update in ins_strs[1]: statements.append(update) # Gather options format = options.get("format", "sql") no_headers = options.get("no_headers", False) quiet = options.get("quiet", False) import_type = options.get("import_type", "definitions") single = options.get("single", True) dryrun = options.get("dryrun", False) do_drop = options.get("do_drop", False) skip_blobs = options.get("skip_blobs", False) skip_gtid = options.get("skip_gtid", False) # Attempt to connect to the destination server conn_options = { 'quiet' : quiet, 'version' : "5.1.30", } servers = connect_servers(dest_val, None, conn_options) destination = servers[0] # Check storage engines check_engine_options(destination, options.get("new_engine", None), options.get("def_engine", None), False, options.get("quiet", False)) if not quiet: if import_type == "both": str = "definitions and data" else: str = import_type print "# Importing %s from %s." % (str, file_name) # Setup variables we will need skip_header = not no_headers if format == "sql": skip_header = False get_db = True check_privileges = False db_name = None file = open(file_name) columns = [] read_columns = False table_rows = [] obj_type = "" definitions = [] statements = [] table_col_list = [] tbl_name = "" skip_rpl = options.get("skip_rpl", False) gtid_command_found = False supports_gtid = servers[0].supports_gtid() == 'ON' skip_gtid_warning_printed = False gtid_version_checked = False # Read the file one object/definition group at a time for row in read_next(file, format): # Check for replication command if row[0] == "RPL_COMMAND": if not skip_rpl: statements.append(row[1]) continue if row[0] == "GTID_COMMAND": gtid_command_found = True if not supports_gtid: # only display warning once if not skip_gtid_warning_printed: print _GTID_SKIP_WARNING skip_gtid_warning_printed = True elif not skip_gtid: if not gtid_version_checked: gtid_version_checked = True # Check GTID version for complete feature support servers[0].check_gtid_version() # Check the gtid_purged value too servers[0].check_gtid_executed("import") statements.append(row[1]) continue # If this is the first pass, get the database name from the file if get_db: if skip_header: skip_header = False else: db_name = _get_db(row) # quote db_name with backticks if needed if db_name and not is_quoted_with_backticks(db_name): db_name = quote_with_backticks(db_name) get_db = False if do_drop and import_type != "data": statements.append("DROP DATABASE IF EXISTS %s;" % db_name) if import_type != "data": if not _skip_object("CREATE_DB", options) and \ not format == 'sql': statements.append("CREATE DATABASE %s;" % db_name) # This is the first time through the loop so we must # check user permissions on source for all databases if db_name is not None: dest_db = Database(destination, db_name) # Make a dictionary of the options access_options = options.copy() dest_db.check_write_access(dest_val['user'], dest_val['host'], access_options) # Now check to see if we want definitions, data, or both: if row[0] == "sql" or row[0] in _DEFINITION_LIST: if format != "sql" and len(row[1]) == 1: raise UtilError("Cannot read an import file generated with " "--display=NAMES") if import_type in ("definitions", "both"): if format == "sql": statements.append(row[1]) else: if obj_type == "": obj_type = row[0] if obj_type != row[0]: if len(definitions) > 0: _process_definitions(statements, table_col_list, db_name) obj_type = row[0] definitions = [] if not _skip_object(row[0], options): definitions.append(row[1]) else: # see if there are any definitions to process if len(definitions) > 0: _process_definitions(statements, table_col_list, db_name) definitions = [] if import_type in ("data", "both"): if _skip_object("DATA", options): continue # skip data elif format == "sql": statements.append(row[1]) else: if row[0] == "BEGIN_DATA": # Start of table so first row is columns. if len(table_rows) > 0: _process_data(tbl_name, statements, columns, table_col_list, table_rows, skip_blobs) table_rows = [] read_columns = True tbl_name = row[1] if not is_quoted_with_backticks(tbl_name): db, sep, tbl = tbl_name.partition('.') q_db = quote_with_backticks(db) q_tbl = quote_with_backticks(tbl) tbl_name = ".".join([q_db, q_tbl]) else: if read_columns: columns = row[1] read_columns = False else: if not single: table_rows.append(row[1]) else: str = _build_insert_data(columns, tbl_name, row[1]) statements.append(str) # Process remaining definitions if len(definitions) > 0: _process_definitions(statements, table_col_list, db_name) definitions = [] # Process remaining data rows if len(table_rows) > 0: _process_data(tbl_name, statements, columns, table_col_list, table_rows, skip_blobs) table_rows = [] # Now process the statements _exec_statements(statements, destination, format, options, dryrun) file.close() # Check gtid process if supports_gtid and not gtid_command_found: print _GTID_MISSING_WARNING if not quiet: print "#...done." return True