#DBHangOps!

Talk shop and learn about MySQL and occasionally some other stores!

Topics and RecordingsLearn From Previous Discussions!

How to JoinSome simple setup help!

Follow Us On

Twitter

#DBHangOps 04/30/14 -- Database Defense (part 2) and 5.6 upgrade stories!

Check out the recording below:

Hello everybody!

Join in #DBHangOps this Wednesday, April, 30, 2014 at 12:00pm pacific (19:00 GMT), to participate in the discussion about:

  • Database Defense part 2
    • Protecting production
    • Security!
  • 5.6 Upgrade stories from Ike Walker

Be sure to check out the #DBHangOps twitter search, the @DBHangOps twitter feed, or this blog post to get a link for the google hangout on Wednesday!

Also hit up the scheduling doodle at http://doodle.com/vbhupvveihzt95z8 and let us know what time works best for you to make it to #DBHangOps!

See all of you on Wednesday!

Show notes

Database Defense (part 2)

  • Security!
  • Shared resource defense (e.g. running in the cloud or using shared virtualized environment)
    • AWS has provisioned IOPS which can help you get more dedicated resources
    • When in shared environments, "over-provisioning" may get you a more dedicated resource
  • Tools/suggestions
    • Server-level changes
      • MAX CONNECTIONS -- you can set this on a server-level, user-leve, and per-user level. If you have multiple applications, assigning them different user and setting a MAX CONNECTION on the user acocunt can help
      • wait_timeout -- influences how long sleeping connections are allowed on the server. Default is 8 hours. Consider setting this lower (perhaps 5-15 minutes)
      • lock_wait_timeout -- This influences how long a query will wait for a lock to be granted before proceeding. The default of this is 1 year. This value does not influence locking behavior in InnoDB (see innodb_lock_wait_timeout)
        • This is checked by some tools to influence their behavior (e.g. pt-online-schema-change)
      • slave_net_timeout -- cross geo replication (cross-continent) and network delays hit 100ms+, replication might constantly stop
    • After you're in a bad state
      • pt-kill -- Get familiar with the tool. It can be supremely helpful during problems!

5.6 Upgrade stories (Ike Walker)