[ https://issues.redhat.com/browse/JBJCA-1399?page=com.atlassian.jira.plugi... ] Flavia Rainone commented on JBJCA-1399: --------------------------------------- So, I'm not sure how this translates to a race condition yet, but the problem is that SemaphoreConcurrentLinkedDequeManagedConnectionPool.increaseCapacity is faulty when it reaches the limit. If it reaches the limit, it doesn't decrement the poolSize when removing the new connection. That leads to a "leak" in the pool size. I.e., the pool will appear bigger than it is. At every interaction, the pool is decremented because the size increases constantly at every attempt to fill the pool. I'm still investigating this [~stephen.fikes], because I want to be sure why it is a race condition before fixing it, since I wasn't able to reproduce the bug in my machine. -- This message was sent by Atlassian Jira (v7.13.8#713008)