On 24 February 2010 03:58, Pratik Bhatt <[email protected]> wrote: > Hi Rafal, > > Thanks for the reply. Actually we are using MS SQL Server 2005 as the back > end, and seeing this issue. Also we are running it on JBoss 5.0GA, with > hibernate provided with jboss. > > Do let me know if you require more information.
OK, so it's a bug. Could you report it on Jira? Then we'll continue discussion there. It would be great if you could do thread dump (kill -3 JVM_PID on unix) while threads are waiting. Additionally, please provide indexes, which you used for ODE_JOB table. > > Regards > Pratik > > On Tue, Feb 23, 2010 at 12:56 AM, Rafal Rusin <[email protected]> wrote: > >> On 22 February 2010 23:22, Pratik Bhatt <[email protected]> wrote: >> > Hello Everyone, >> > >> > I just wanted to check, if there was any way to disable or defer the >> > deletion of entries from the ode_jobs table, maybe through a >> configuration >> > setting. >> > >> > From what I understand, the deletion of records is more of housekeeping, >> and >> > not something that needs to happen immediately. I am enquiring about it >> > because, the delete queries are causing a lot of deadlocks anytime we run >> a >> > large number of threads in parallel. >> > >> > Thanks for the help in advance. >> >> There is no way to configure such behaviour. If you really want to do >> this, you can modify >> >> scheduler-simple/src/main/java/org/apache/ode/scheduler/simple/JdbcDelegate.java. >> But could you provide your configuration? Is it Derby embedded >> database? If so, it's known problem and we are working on migrating to >> H2 database to resolve this >> (http://issues.apache.org/jira/browse/ODE-666). You can switch to >> MySQL or Oracle as well. This should help. >> >> Regards, >> -- >> Rafał Rusin >> http://rrusin.blogspot.com >> http://www.touk.pl >> http://top.touk.pl >> > Regards, -- Rafał Rusin http://rrusin.blogspot.com http://www.touk.pl http://top.touk.pl
