[ 
https://issues.apache.org/jira/browse/SYNCOPE-460?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13859954#comment-13859954
 ] 

Guido Wimmel commented on SYNCOPE-460:
--------------------------------------

I've investigated a bit further into the test failure 
ReportTestITCase.executeAndExport() if the patch is applied. The following 
seems to happen:
1. ReportController.execute() calls Quartz Scheduler.registerJob() and 
.triggerJob(), which signal QuartzSchedulerThread
2. QuartzSchedulerThread wakes up several times but does not find the created 
trigger; then it waits for its usual idle wait time (approximately 30s)
3. after this time, QuartzSchedulerThread wakes up again, finds the trigger 
now, and executes ReportJob

Without the patch, QuartzSchedulerThread finds the trigger when waking up and 
executes ReportJob immediately.

Maybe some problem with synchronized/transactional access to the Quartz 
database surfaces when the connection pool is used. I'm a bit lost here 
currently - maybe someone familiar with Quartz internals has an idea, or 
related problems have already been encountered sometime?



> Default datasource uses BasicDataSource
> ---------------------------------------
>
>                 Key: SYNCOPE-460
>                 URL: https://issues.apache.org/jira/browse/SYNCOPE-460
>             Project: Syncope
>          Issue Type: Improvement
>          Components: core
>    Affects Versions: 1.1.5
>            Reporter: Oliver Wulff
>            Assignee: Francesco Chicchiriccò
>            Priority: Minor
>             Fix For: 1.1.6, 1.2.0
>
>         Attachments: patch.git, syncope-460_20121220.patch
>
>
> Replace org.springframework.jdbc.datasource.DriverManagerDataSource by 
> org.apache.commons.dbcp.BasicDataSource as recommended by spring:
> http://docs.spring.io/spring/docs/current/spring-framework-reference/html/jdbc.html#jdbc-connections
> Make property "isolationLevelName" configurable in persistence.properties 
> with default ISOLATION_READ_COMMITTED
> See full discussion in dev@ ML: http://markmail.org/message/33aheawb5irzl3xc



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

Reply via email to