[
https://issues.apache.org/jira/browse/POOL-259?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Phil Steitz updated POOL-259:
-----------------------------
Description:
The local variable waitTime used by borrowObject in both GOP and GKOP is not
updated unless borrowMaxWaitMillis is non-negative and blockWhenExhausted is
true. This effectively zeros the JMX client wait time properties for these
configurations. Even when there is a maxWait configured and the pool is set to
block when exhausted, what ends up being reported is only the wait time on the
instance queue. Note that this bug is JMX reporting only.
I am inclined to fix this by setting waitTime at the beginning and updating it
at the end (including activate and validate times). We should also add test
cases for this and the other stats exposed via JMX. Patches welcome, as it may
take me a little while to get to this.
was:The local variable waitTime used by borrowObject in both GOP and GKOP is
not updated unless borrowMaxWaitMillis is non-negative and blockeWhenExhausted
is true. This effectively zeros the JMX client wait time properties for these
configurations. Even when there is a maxWait configured and the pool is set to
block when exhausted, what ends up being reported is only the wait time on the
instance queue. I am inclined to fix this by setting waitTime at the beginning
and updating it at the end (including activate and validate times). We should
also add test cases for this and the other stats exposed via JMX. Patches
welcome, as it may take me a little while to get to this.
> Client wait time is not updated when max wait is configured to be unlimited
> ---------------------------------------------------------------------------
>
> Key: POOL-259
> URL: https://issues.apache.org/jira/browse/POOL-259
> Project: Commons Pool
> Issue Type: Bug
> Affects Versions: 2.2
> Reporter: Phil Steitz
>
> The local variable waitTime used by borrowObject in both GOP and GKOP is not
> updated unless borrowMaxWaitMillis is non-negative and blockWhenExhausted is
> true. This effectively zeros the JMX client wait time properties for these
> configurations. Even when there is a maxWait configured and the pool is set
> to block when exhausted, what ends up being reported is only the wait time on
> the instance queue. Note that this bug is JMX reporting only.
> I am inclined to fix this by setting waitTime at the beginning and updating
> it at the end (including activate and validate times). We should also add
> test cases for this and the other stats exposed via JMX. Patches welcome, as
> it may take me a little while to get to this.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
