Supporting continuity and open collaboration

A few interesting findings on MariaDB and MySQL scalability, multi-table OLTP RO

It’s been almost a year since I benchmarked MariaDB and MySQL on our good old 4 CPU / 32 Cores / 64 Threads Sandy Bridge server. There seem to be a few interesting things happened since that time.

image

  • MySQL 5.6.23 peak throughput dropped by ~8% compared to 5.6.14. Looks like this regression appeared in MySQL 5.6.21.
  • 10.0.18 (git snapshot) peak threads increased by ~20% compared to 10.0.9 and reached parity with 5.6.23 (not with 5.6.20 though).
  • 10.1.4 (git snapshot) and 5.7.5 are the champions (though 10.1.4 was usually 1-5% faster). Both have similar peaks @ 64 threads. They seem to be pushing this system limits, since both have capacity to scale more.
  • 5.7.6 has serious (~50%) scalability regression compared to 5.7.5. There is heavy LOCK_plugin mutex contention, which seem to be caused by query rewrite plugin framework. I couldn’t find way to disable it. Hope this will be fixed in coming releases.

Raw data available at google doc.

1 Comment

  1. 2015-03-26    

    The results are nice.

    LOCK_plugin, causing odd segfaults and mutex contention since 2008. I wonder how frequently it is used for things that really are plugins and loaded dynamically.

    A few good ones…
    http://bugs.mysql.com/bug.php?id=70218
    https://bugs.mysql.com/bug.php?id=34409
    https://bugs.mysql.com/bug.php?id=72215

Platinum Sponsors

MariaDB Foundation Platinum sponsors

Thank you, Booking.com! Thank you, Alibaba Cloud! Thank you, Tencent Cloud! Thank you, Microsoft! Thank you, MariaDB Corporation!

Gold Sponsors

MariaDB Foundation Gold sponsors

Thank you, DBS! Thank you, Visma! Thank you, IBM! Thank you, Tencent Games!

Code statistics