Skip to content
  • Sujatha Sivakumar's avatar
    0327e99a
    Bug#28384862: PB2 RPL_SERVER_UUID.TEST IS FAILING · 0327e99a
    Sujatha Sivakumar authored
    SPORADICALLY ON WEEKLY/DAILY 5.6/5.7
    
    Problem:
    =======
    In test case 4, while waiting for slave threads to go down
    an error condition is reached and stop_slave.inc fails with
    following error.
    
    included from ./include/stop_slave.inc at line 84:
    At line 142: Error condition reached in
    include/wait_for_slave_param.inc
    
    Analysis:
    ========
    Prior to test case 4 master's UUID file is temporarily
    replaced with slave's UUID file to test some error
    scenarios.  Once all these tests are done master's original
    UUID file is restored and the master server is restarted.
    
    As part of test case 4 START SLAVE IO_THREAD is invoked and
    an assert is added which verifies that slave is still
    referring to master's original UUID.
    
    Upon successful test stop_slave_io.inc is called to stop both
    threads.
    
    On slow platforms the START SLAVE IO_THREAD tries to connect
    to master server before the master server is ready for
    connections.
    
    Test failure debug log shows like this:-
    ======================================
    **** ERROR: Last_IO_Errno = '2013' while waiting for slave
    parameter Slave_IO_Running = No ****
    DO_NOT_CHECK_IN_THIS_LINE: include/show_rpl_debug_info.inc
    should only be used for debugging. Never check in a test
    that calls it on success.  CURRENT_CONNECTION = 'slave'
    
    ############################## slave ######################
    
    **** SHOW WARNINGS on slave ****
    SHOW WARNINGS;
    Level   Code    Message
    
    **** SELECT replication-related variables on slave ****
    SELECT NOW(), @@SERVER_ID, @@SERVER_UUID, @@PORT;
    NOW()   2018-07-13 07:41:54
    
    Slave gave up at 07:41:54
    
    Last_IO_Errno   2013
    Last_IO_Error   error connecting to master
    'root@127.0.0.1:13050' - retry-time: 1  retries: 1
    
    Master server's error log shows:-
    ===============================
    2018-07-13 07:41:54 32093 [Note] Event Scheduler: Loaded 0
    events
    2018-07-13 07:41:54 32093 [Note]
    /export/home/pb2/test/sb_2-29588355-1531456600.5/mysql/usr/
    sbin/mysqld-debug: ready for connections.
    
    IO thread tried to connect master just before master server
    was up and it got an error 2013. "stop_slave.inc" was
    invoked before the IO thread retried and hence test fails.
    
    Fix:
    ===
    Replaced START SLAVE IO_THREAD with start_slave_io.inc.
    0327e99a
    Bug#28384862: PB2 RPL_SERVER_UUID.TEST IS FAILING
    Sujatha Sivakumar authored
    SPORADICALLY ON WEEKLY/DAILY 5.6/5.7
    
    Problem:
    =======
    In test case 4, while waiting for slave threads to go down
    an error condition is reached and stop_slave.inc fails with
    following error.
    
    included from ./include/stop_slave.inc at line 84:
    At line 142: Error condition reached in
    include/wait_for_slave_param.inc
    
    Analysis:
    ========
    Prior to test case 4 master's UUID file is temporarily
    replaced with slave's UUID file to test some error
    scenarios.  Once all these tests are done master's original
    UUID file is restored and the master server is restarted.
    
    As part of test case 4 START SLAVE IO_THREAD is invoked and
    an assert is added which verifies that slave is still
    referring to master's original UUID.
    
    Upon successful test stop_slave_io.inc is called to stop both
    threads.
    
    On slow platforms the START SLAVE IO_THREAD tries to connect
    to master server before the master server is ready for
    connections.
    
    Test failure debug log shows like this:-
    ======================================
    **** ERROR: Last_IO_Errno = '2013' while waiting for slave
    parameter Slave_IO_Running = No ****
    DO_NOT_CHECK_IN_THIS_LINE: include/show_rpl_debug_info.inc
    should only be used for debugging. Never check in a test
    that calls it on success.  CURRENT_CONNECTION = 'slave'
    
    ############################## slave ######################
    
    **** SHOW WARNINGS on slave ****
    SHOW WARNINGS;
    Level   Code    Message
    
    **** SELECT replication-related variables on slave ****
    SELECT NOW(), @@SERVER_ID, @@SERVER_UUID, @@PORT;
    NOW()   2018-07-13 07:41:54
    
    Slave gave up at 07:41:54
    
    Last_IO_Errno   2013
    Last_IO_Error   error connecting to master
    'root@127.0.0.1:13050' - retry-time: 1  retries: 1
    
    Master server's error log shows:-
    ===============================
    2018-07-13 07:41:54 32093 [Note] Event Scheduler: Loaded 0
    events
    2018-07-13 07:41:54 32093 [Note]
    /export/home/pb2/test/sb_2-29588355-1531456600.5/mysql/usr/
    sbin/mysqld-debug: ready for connections.
    
    IO thread tried to connect master just before master server
    was up and it got an error 2013. "stop_slave.inc" was
    invoked before the IO thread retried and hence test fails.
    
    Fix:
    ===
    Replaced START SLAVE IO_THREAD with start_slave_io.inc.
Loading