Bug #50176 rpl_backup_block times out on mysql-backup-backport
Submitted: 8 Jan 2010 10:08 Modified: 13 Jan 2010 18:29
Reporter: Andrei Elkin Email Updates:
Status: Patch queued Impact on me:
None 
Category:MySQL Server: Backup Severity:S3 (Non-critical)
Version: OS:Any
Assigned to: Assigned Account CPU Architecture:Any
Triage: Triaged: D3 (Medium)

[8 Jan 2010 10:08] Andrei Elkin
Description:
A recent failure with the test

https://central.sun.net/http://pb2.norway.sun.com/?template=mysql_show_test_failure&test_f...

deals with the wait-for-slave-io-to-stop timeout elapsed

@@ -119,30 +119,38 @@
 SET @@global.backupdir= @@global.datadir;
 # Now verify slave cannot start.
 START SLAVE IO_THREAD;
+**** ERROR: timeout after 3000 deci-seconds while waiting for slave parameter Slave_IO_Running = No ****
+Message: Failed while waiting for slave IO thread to stop
+Current connection is 'slave2'
+Note: the following output may have changed since the failure was detected

It's not clear yet whether the failure is about the test or the server.

Notice, that Bug #47699 although happened with the test still contains different
symptoms (resembling Bug #50175)

How to repeat:
To see PB2 or run the test.
[11 Jan 2010 10:34] Bugs System
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/96475

3050 Andrei Elkin	2010-01-11
      Bug #50176 rpl_backup_block times out on mysql-backup-backport
      
      The timeout happened as a combination of actual values for the global master_retry_connect
      and mi->connect_retry set way too big and treating ER_MASTER_BLOCKING_SLAVES as a reconnectable
      error.
      
      Fixed with setting the two parameters of the 2nd slave to correspond to mtr defaults
      and making `ER_MASTER_BLOCKING_SLAVES' critical so that auto-reconnecting is not to happen.
     @ mysql-test/suite/rpl/r/rpl_backup_block.result
        results changed.
     @ mysql-test/suite/rpl/t/rpl_backup_block.cnf
        Setting the slave connect retry to the mtr's default.
     @ mysql-test/suite/rpl/t/rpl_backup_block.test
        setting up MASTER_CONNECT_RETRY=1 for the 2nd slave explicitly to the value of 
        master-slave-reset.inc;
        simplifying output in the case 2;
        correcting the suppressions.
     @ sql/slave.cc
        Making `ER_MASTER_BLOCKING_SLAVES' critical so that auto-reconnecting is not to happen.
[12 Jan 2010 15:50] Bugs System
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/96626

3050 Andrei Elkin	2010-01-12
      Bug #50176 rpl_backup_block times out on mysql-backup-backport
      
      The timeout happened as a combination of actual values for the global master_retry_connect
      and mi->connect_retry set way too big and treating ER_MASTER_BLOCKING_SLAVES as a reconnectable
      error.
      
      Fixed with setting the two parameters of the 2nd slave to correspond to mtr defaults
      and making `ER_MASTER_BLOCKING_SLAVES' critical so that auto-reconnecting is not to happen.
     @ mysql-test/suite/rpl/r/rpl_backup_block.result
        results changed.
     @ mysql-test/suite/rpl/t/rpl_backup_block.cnf
        Setting the slave connect retry to the mtr's default.
     @ mysql-test/suite/rpl/t/rpl_backup_block.test
        setting up MASTER_CONNECT_RETRY=1 for the 2nd slave explicitly to the value of 
        master-slave-reset.inc;
        simplifying output in the case 2;
        correcting the suppressions.
     @ sql/share/errmsg-utf8.txt
        editing ER_MASTER_BLOCKING_SLAVES error message.
     @ sql/share/errmsg.txt
        editing ER_MASTER_BLOCKING_SLAVES error message.
     @ sql/slave.cc
        Making `ER_MASTER_BLOCKING_SLAVES' critical so that auto-reconnecting is not to happen.
[13 Jan 2010 18:29] Andrei Elkin
Pushed to backup-port.