[jira] Updated: (SANDESHA2-184) Preventing Dead lock isses with jdbc persistence storage

     [ https://issues.apache.org/jira/browse/SANDESHA2-184?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Amila Chinthaka Suriarachchi updated SANDESHA2-184:
---------------------------------------------------

    Attachment: patch.txt

This patch implement the above mentioned improvement.
This also includes a change to MakeConnectionProcessor.java to prevent a deadlock which can occur if a thread does not commit before getting the worker lock.

> Preventing Dead lock isses with jdbc persistence storage
> --------------------------------------------------------
>
>                 Key: SANDESHA2-184
>                 URL: https://issues.apache.org/jira/browse/SANDESHA2-184
>             Project: Sandesha2
>          Issue Type: Improvement
>            Reporter: Amila Chinthaka Suriarachchi
>         Attachments: patch.txt
>
>
> There are some random deadlocks occurs with jdbc persistence storage since Sandesha2 core access the table resources in different orders. Although these deadlocks are rear there is a possibility of getting them. 
> As a solution for this problem we can lock the transactions at the storage level and let only one transaction to proceed. Then there is no chance for dead locks.
> This appears to be not a proper solution but it can prevent the deadlock occurrences.
> This may reduce the performance since it prevents parallel processing and is not solving the problem in a distributed environment.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.