Percona Server 5.1.52-12.3

Percona is glad to announce the release of Percona Server 5.1.52-12.3 on December 13th, 2010 (Downloads are available here and from the Percona Software Repositories).

Functionality Added or Changed

  • Percona Server 5.1.52-12.3 is now based on MySQL 5.1.52.
  • New Features Added:
    • HandlerSocket plugin support has been added to all RPM and Debian packages. This is an experimental feature. (Inada Naoki, DeNA Co., Ltd.)
    • Lock-Free SHOW SLAVE STATUS - Allows examination of slave status even when a lock prevents SHOW SLAVE STATUS from operating. (Oleg Tsarev)
    • Fast Shutdown - Sleeping threads could cause a delay of up to 10 seconds during InnoDB shutdown. The sleeping state of threads was made interruptible to avoid this. (contributed by Kristian Nielsen) (Alexey Kopytov)
  • Other Changes: None

Bugs Fixed

  • Bug #640576 - The false error “innodb_extra_undoslots option is disabled but it was enabled before” was sometimes reported after upgrading from MySQL 5.0. (Yasufumi Kinoshita)
  • Bug #643463 - Shutting down XtraDB could take up to 10 seconds, even when there was no actual work to do. (Fix contributed by Kristian Nielsen) (Alexey Kopytov)
  • Bug #663757 - On the FreeBSD platform, the “gcc atomic built-in” function isn’t provided, so response_time_distribution now uses the native atomic API on FreeBSD. (Oleg Tsarev)
  • Bug #673426 - Use of some system variables as command-line options caused a crash or undefined behavior. (Oleg Tsarev)
  • Bug #673562 - Debug build was broken due a to failing compile-time assertion in (Alexey Kopytov)
  • Bug #673567 - Compiler could produce spurious warnings when building on non-Linux platforms. A check is now made to see if clock_gettime() is present in librt at the configure stage. If yes, `` -lrt`` is added to LIBS. (Alexey Kopytov)
  • Bug #673929 - Query cache misses were being reported for some queries when hits were actually occurring. (Oleg Tsarev)
  • Bug #676146 - The development environment test of log_slow_verbosity=innodb on a slave for row-based replication was not working correctly. (Oleg Tsarev)
  • Bug #676147 - The development environment test of option log_slow_slave_statements for row-based replication was not working correctly. (Oleg Tsarev)
  • Bug #676148 - Similar to Bug #676147. A check is now made for the replication type to test. (Oleg Tsarev)