You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@commons.apache.org by "Giambattista Bloisi (JIRA)" <ji...@apache.org> on 2009/07/02 14:11:47 UTC
[jira] Created: (POOL-147) Thread stuck in GenericObjectPool
borrowObject()
Thread stuck in GenericObjectPool borrowObject()
------------------------------------------------
Key: POOL-147
URL: https://issues.apache.org/jira/browse/POOL-147
Project: Commons Pool
Issue Type: Bug
Affects Versions: 1.5.1
Reporter: Giambattista Bloisi
Priority: Minor
In my application I found a thread stuck at GenericObjectPool.borrowObject(GenericObjectPool.java:1099) even though resources were availbale (in my case connections for DBCP).
After an analysis of the code I think I spotted what could have caused this behavior. GenericObjectPool.java:1099 line corresponds to an indefinite wait to a "latch" object. This wait should be awaken when new resources are made available to the pool. The problem is likely that borrowObject enters in wait without performing "latch.getPair()" check in a critical section.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (POOL-147) Thread stuck in GenericObjectPool
borrowObject()
Posted by "Giambattista Bloisi (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/POOL-147?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Giambattista Bloisi updated POOL-147:
-------------------------------------
Attachment: 01-borrowObject.patch
Proposed patch
> Thread stuck in GenericObjectPool borrowObject()
> ------------------------------------------------
>
> Key: POOL-147
> URL: https://issues.apache.org/jira/browse/POOL-147
> Project: Commons Pool
> Issue Type: Bug
> Affects Versions: 1.5.1
> Reporter: Giambattista Bloisi
> Priority: Minor
> Attachments: 01-borrowObject.patch
>
>
> In my application I found a thread stuck at GenericObjectPool.borrowObject(GenericObjectPool.java:1099) even though resources were availbale (in my case connections for DBCP).
> After an analysis of the code I think I spotted what could have caused this behavior. GenericObjectPool.java:1099 line corresponds to an indefinite wait to a "latch" object. This wait should be awaken when new resources are made available to the pool. The problem is likely that borrowObject enters in wait without performing "latch.getPair()" check in a critical section.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (POOL-147) Thread stuck in GenericObjectPool
borrowObject()
Posted by "Mark Thomas (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/POOL-147?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Mark Thomas resolved POOL-147.
------------------------------
Resolution: Fixed
Fix Version/s: 1.5.2
Many thanks for the report, your analysis and the patch. One day all bug reports will be like this one ;)
I only made one change to your patch which was to break out of the loop as soon as we know we have been allocated an object.
Again, many thanks.
> Thread stuck in GenericObjectPool borrowObject()
> ------------------------------------------------
>
> Key: POOL-147
> URL: https://issues.apache.org/jira/browse/POOL-147
> Project: Commons Pool
> Issue Type: Bug
> Affects Versions: 1.5.1
> Reporter: Giambattista Bloisi
> Priority: Minor
> Fix For: 1.5.2
>
> Attachments: 01-borrowObject.patch
>
>
> In my application I found a thread stuck at GenericObjectPool.borrowObject(GenericObjectPool.java:1099) even though resources were availbale (in my case connections for DBCP).
> After an analysis of the code I think I spotted what could have caused this behavior. GenericObjectPool.java:1099 line corresponds to an indefinite wait to a "latch" object. This wait should be awaken when new resources are made available to the pool. The problem is likely that borrowObject enters in wait without performing "latch.getPair()" check in a critical section.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Closed: (POOL-147) Thread stuck in GenericObjectPool
borrowObject()
Posted by "Phil Steitz (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/POOL-147?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Phil Steitz closed POOL-147.
----------------------------
> Thread stuck in GenericObjectPool borrowObject()
> ------------------------------------------------
>
> Key: POOL-147
> URL: https://issues.apache.org/jira/browse/POOL-147
> Project: Commons Pool
> Issue Type: Bug
> Affects Versions: 1.5.1
> Reporter: Giambattista Bloisi
> Priority: Minor
> Fix For: 1.5.2
>
> Attachments: 01-borrowObject.patch
>
>
> In my application I found a thread stuck at GenericObjectPool.borrowObject(GenericObjectPool.java:1099) even though resources were availbale (in my case connections for DBCP).
> After an analysis of the code I think I spotted what could have caused this behavior. GenericObjectPool.java:1099 line corresponds to an indefinite wait to a "latch" object. This wait should be awaken when new resources are made available to the pool. The problem is likely that borrowObject enters in wait without performing "latch.getPair()" check in a critical section.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.