Edit File: wait_slave_status.inc
# include/wait_slave_status.inc # # Created by Matthias Leich # # SUMMARY # # Waits until slave has reached certain state or maximum time reached. # # (This script will not work, when the SHOW command delivers more than one # result record, because only the first record will be caught.) # # USAGE # # Set $result_pattern in test file and source this file: # # let $result_pattern= <pattern used for LIKE on the result of # SHOW STATUS SLAVE> # --include wait_slave_status.inc # # EXAMPLE # # The script rpl_until.test: # ... # --replace_result $MASTER_MYPORT MASTER_MYPORT # --replace_column 1 # 9 # 23 # 33 # # --vertical_results show slave status; # # outputs # show slave status; # Slave_IO_State # # Master_Host 127.0.0.1 # Master_User root # Master_Port MASTER_MYPORT # Connect_Retry 1 # Master_Log_File master-bin.000001 # Read_Master_Log_Pos 776 # Relay_Log_File slave-relay-bin.000004 # Relay_Log_Pos # # Relay_Master_Log_File master-bin.000001 # Slave_IO_Running Yes # Slave_SQL_Running No # Replicate_Do_DB # Replicate_Ignore_DB # Replicate_Do_Table # Replicate_Ignore_Table # Replicate_Wild_Do_Table # Replicate_Wild_Ignore_Table # Last_Errno 0 # Last_Error # Skip_Counter 0 # Exec_Master_Log_Pos 319 # Relay_Log_Space # # Until_Condition Master # Until_Log_File master-bin.000001 # Until_Log_Pos 319 # Master_SSL_Allowed No # Master_SSL_CA_File # Master_SSL_CA_Path # Master_SSL_Cert # Master_SSL_Cipher # Master_SSL_Key # Seconds_Behind_Master # # # The main problem with the "show slave status;" in rpl_until is, that # depending on the total test engine power and the current load caused by # other processes, the expected slave status might be not reached though # it will happen in maybe some seconds. # # The typical problem with rpl_until is that Slave_IO_Running is "No" # instead of "Yes". # # The expected result follows the LIKE pattern: # # let $result_pattern= '%127.0.0.1%root%1%master-bin.000001%776%slave-relay-bin.000004%master-bin.000001%Yes%No%0%0%319%Master%master-bin.000001%319%No%'; # # The Slave_IO_Running value is the "Yes" just after the "master-bin.000001". # # How to get this pattern ? # # Any lines "--replace_result ..." and "--replace_colum ..." just before # the SHOW TABLE STATUS and of course the expected result itself # show us columns where the content must be unified, because it is non # deterministic or it depends on the current test environment. # # Unfortunately "--replace_result ..." and "--replace_colum ..." do not # affect the result of our assignment let $my_val= `SHOW SLAVE STATUS`; # Therefore such content must be covered by '%'. # # Please be careful. A more simple pattern might be dangerous, because we # might get "wrong" matches. Example: There might be several "Yes" and "No" # within one result row. # ############################################################################### # We do not want to print the auxiliary commands, because they are not of # interest and their amount will vary depending how fast we get the # desired state. --disable_query_log # The protocol should show # - the setting of $result_pattern and # - that this file is sourced , # because this increases the chance to use the protocol as replay script. eval SELECT "let \$result_pattern= $result_pattern ;" AS ""; SELECT '--source include/wait_slave_status.inc' AS ""; let $show_type= SLAVE STATUS; let $show_pattern= $result_pattern; --enable_query_log -- source $MYSQL_TEST_DIR/suite/engines/funcs/t/wait_show_pattern.inc if (!$success) { let $message= ! Attention: Timeout in wait_slave_status.inc. | Possible reasons with decreasing probability: | - The LIKE pattern is wrong, because the | testcase was altered or the layout of the | SHOW SLAVE STATUS result set changed. | - There is a new bug within the replication. | - We met an extreme testing environment and timeout is | too small.; --source include/show_msg80.inc --echo DEBUG INFO START (wait_slave_status.inc): --echo $result_pattern --vertical_results show slave status; --echo DEBUG INFO END }
Back to File Manager