org.apache.ode.bpel.memdao.ProcessDaoImpl cleans instances which are already
cleaned and the instances are nulls
----------------------------------------------------------------------------------------------------------------
Key: ODE-678
URL: https://issues.apache.org/jira/browse/ODE-678
Project: ODE
Issue Type: Bug
Components: BPEL Runtime, JBI Integration
Affects Versions: 1.2, 1.3.4
Environment: WinXP, Fuse 3.4.0.4, ODE 1.2 JBI
Reporter: Mateusz Nowakowski
Fix For: 1.3.4
I've got such logs:
2009-10-14 06:39:35,234 [pool-5-thread-21] DEBUG
org.apache.ode.bpel.memdao.ProcessDaoImpl - Removing completed process instance
597 from in-memory store.
(...)
2009-10-14 06:50:47,412 [pool-5-thread-48] WARN
org.apache.ode.bpel.memdao.ProcessDaoImpl - Discarding in-memory instance 597
because it exceeded its time-to-live: null
It occurs for valid requests which don't generate any underlying invokes.
So the process consists only of receive and reply activities (and several ifs,
but all return false and its contents are not executed).
The whole transactions are completed successfully.
What is more:
The cleaning task is not a periodic task!, it is invoked only in:
public ProcessInstanceDAO createInstance(CorrelatorDAO correlator)
so only when new request comes to ODE.
I think the bug occurs in the latest 1.x revision as well, because I haven't
noticed any change in instanceCompleted(..) method .
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
