MySQL 5.6.15 Community Release Notes

Thank you to the MySQL Community, on behalf of the MySQL team @ Oracle. Your bug reports, testcases and patches have helped create a better MySQL 5.6.15.

In particular:

  • Thanks to Yoshinori Matsunobu, who reported that the performance_schema did not instrument locks related to semisync replication. Bug #70577.
  • Thanks to Sebastian Strzelczyk, who reported excessive memory usage when querying INFORMATION_SCHEMA.INNODB_FT_INDEX. Sebastian uploaded a number of reports of memory usage, which was helped us generate a good reproducible testcase. Bug #70329.
  • Thanks to Inaam Rana, who suggested improvements to InnoDB counters in information_schema, and contributed a patch! Bug #70241.
  • Thanks to Yoshinori Matsunobu, who reported inefficiency in InnoDB memcached API. Yoshinori’s feedback was also part of the 1 million QPS record set for 5.7. Bug #70172.
  • Thanks to Nizameddin Ordulu, who identified a case where crash recovery might not be possible with InnoDB. Nizameddin pointed out where in the code the bug existed, and we were quickly able to identify that this was a regression from an earlier bug fix. Bug #69122.
  • Thanks to Elena Stepanova, who identified a case where SELECT COUNT queries would run very slowly when run concurrently with a “LOAD DATA” operation. Elena also provided a testcase! Bug #69141.
  • Thanks to Daniël van Eeden, who reported a confusing/incorrect error message when using the InnoDB memcached API. Bug #68684.
  • Thanks to Monty Widenius and Michal Sulik, who both reported an issue where a unique composite index would return wrong results. Bug #70038, Bug #60220.
  • Thanks to Edward Dore, who reported an issue when converting tables from MyISAM to InnoDB. It was from a table that Edward provided that we were able to reproduce the issue! Bug #50366.
  • Thanks to Elena Stepanova who reported an error when deleting from a partitioned table, and provided a testcase! Bug #67982.
  • Thanks to Santosh Praneeth Banda for reporting a security issue. Bug #70349.
  • Thanks to Vlad Lesin, who reported that certain LOAD DATA statements would appear in the binary log twice, and provided a testcase! Bug #70277.
  • Thanks to Yoshinori Matsunobu, who reported that transactions could be lost when starting/stopping slaves with GTIDs. Yoshinori also provided a small program which we were able to use to reproduce the issue. Bug #69943.
  • Thanks to raolh rao for reporting a security issue. #70045.
  • Thanks to Miguel Angel Nieto, who reported that LAST_INSERT_ID is replicated incorrectly when replication filters are used. #69861.
  • Thanks to Santosh Praneeth Banda, who reported an issue with GTIDs and a patch with a suggested fix! Bug #69618.
  • Thanks to Maciej Dobrzanski, who reported an issue where replication would break following invalid grant statements. Bug #68892.
  • Thanks to Rich Prohaska and Przemyslaw Malkowski, who both reported a problem with BINARY columns and online DDL. Bugs #67141 and #69580.
  • Thanks to Jan Staněk, who submitted a patch with code improvements following the result of a Coverity scan. Bug #68918.
  • Thanks to Alexey Kopytov, who reported an issue in debug builds of MySQL, Bug #69653.
  • Thanks to Arthur O’Dwyer, who pointed out that COUNT DISTINCT would incorrectly count NULL rows. Bug #69841.
  • Thanks to Lance Ivy for discovering an issue where UPDATE does not update rows in the correct order. Bug #68656.
  • Thanks to Elena Stepanova for pointing out that mysql_upgrade will fail with the InnoDB storage engine disabled. Bug #70152.
  • Thanks to Arthur O’Dwyer, who reported that prefix indexes on text columns could return wrong results. Bug #70341.
  • Thanks to Yoshinori Matsunobu who pointed out that PERFORMANCE_SCHEMA overhead was very high in frequent connect/disconect scenarios. Bug #70018.
  • Thanks to Moritz Schuepp, who reported that host_cache_size set in a configuration file was ignored. Bug #70552.
  • Thanks to Saverio Miroddi, who reported an issue with InnoDB FULLTEXT and using savepoints. Bug #70333.
  • Thanks to Hartmut Holzgraefe, who reported that users with a legacy hostname may not be possible to drop. Hartmut also provided a patch! Bug #62255, #62254.
  • Thanks to Honza Horak, who reported that MySQL incorrectly assumed all ethernet interfaces start with ethX. Honza also provided a patch! Bug #63055.
  • Thanks to Honza Horak, who reported a number of code improvements (and patches) resulting from Coverity analysis. Bug #70591.
  • Thanks to Jan Staněk, who suggested code improvements (and a patch) resulting from Coverity analysis. Bug #68896.
  • Thanks to Roderik van der Veer, who reported that MySQL would not compile on Mac OS X 10.9. Thanks also to David Peixotto who provided an example patch of how to fix the problem. While we didn’t end up using David’s patch, we certainly value his contribution. Bug #70542.
  • Thanks to Olaf van der Spek and Simon Mudd, who both reported an issue that affected server shutdown. Bugs #18256, #69975.

Thank you again all the names listed above. In particular, I would like to call out the names that appear more than once:

Yoshinori Matsunobu (4), Elena Stepanova (3), Santosh Praneeth Banda (2), Jan Staněk (2), Arthur O’Dwyer (2), Honza Horak (2).

And the names of those who contributed patches:

Inaam Rana, Santosh Praneeth Banda, Hartmut Holzgraefe, Honza Horak, Jan Staněk

If I missed a name here, please let me know!

- Morgan

Published by

morgo

I joined MySQL AB in 2006, left, and am now back at Oracle working on the MySQL team. I’ve also worked at Percona and InPowered.