All hikaricp connection are in use and never recovery util restart

[jilen]

• HikariCP Version: 2.3.7
• Slick: 3.1.0
• mysql-connector-j: 5.1.30

StackOverflow Link
Full log https://gist.github.com/jilen/f4edf44f4013ffe7a1eb

[brettwooldridge]

@jilen Did you get an answer for this in the Slick discussion group?

[jilen]

@brettwooldridge No https://groups.google.com/forum/#!topic/scalaquery/IBipBgCsJOI

[jilen]

Seems related to slick/slick#1274

[kosakriszi]

I have this exact issue, but I am not using slick.

HikariCP: 2.4.3
mysql-connector-j: 5.1.14 (I have no control over this)

Hikari Debug Log: https://gist.github.com/kosakriszi/7bfca3f844bb6f4d7a47

What's interesting to me is that this issue only occurs when executing an SQL insert statement. All other cases seem to be functioning properly.

[jnehlmeier]

@kosakriszi are you sure your code closes connections correctly? AFAICT in HikariCP active means a connection is handed out of the pool and not yet returned to the pool by calling connection.close(). So active does not necessarily mean that SQL statements are currently executed on all these connections.

You can use the leakDetectionThreshold parameter to get stack traces in the log when a connection has been handed out too long compared to your average query duration. That way you can investigate your code to make sure it behaves correctly.

[kosakriszi]

@jnehlmeier Thanks for the reply! I am always sure to close my connections after each query, so I do not believe that that is the issue.

I will try out the leakDetectionThreshold parameter, and see what more information I can gather.

[kosakriszi]

I actually had a rouge connection that I wasn't closing, so my issue is invalidated.