Random hibernate exceptions giving me the 500 page

[biddster]

I've installed a monolithic Stash stack on a single Centos 6.4 x64 virtual machine.

I get the Oops 500 page appearing randomly and I'm pretty sure I'm bumping into this bug.

https://confluence.atlassian.com/display/STASHKB/Connection+Timeout+with+MySQL+Database

The Stash connection pool timeout (default is 10 mins) is greater than the mysql cookbook wait_timeout (default 3 mins).

I've updated my node attributes to this:

node['mysql']['tunable']['wait_timeout']         = '720'

Now Stash reaps connections before mysql has the chance and the errors haven't reappeared for the last 24 hours.

Should we include a setting in stash-config.properties for db.pool.idle.testInterval so that we can override the stash default and default it to node['mysql']['tunable']['wait_timeout'] or 120 seconds where mysql isn't on the box?

[bflad]

Interesting. I hadn't run into this behavior yet. I'll be able to look into this more tomorrow

[bflad]

This is a great find! Having long running idle connections seems generally undesirable. I'm opting for lowering the default Stash database pool idle test interval down to 2 minutes (from 10). This issue will be closed by my next commit and 3.1.2 release.