Hi Alex, I tried with the latest nightly build [build # 86 jan 7 , 2009] http://hudson.zones.apache.org/hudson/view/ODE/job/ODE-trunk/lastStableBuild/
After the service was invoked and was waiting for a response, i killed the ode. After restarting the ode, this what i got on console: Jan 7, 2009 11:57:17 AM org.apache.catalina.startup.Catalina start INFO: Server startup in 54985 ms DEBUG - GeronimoLog.debug(66) | handleWorkEvent: InvokeCheck event for mexid hqejbhcnphr3w17uvkwuxl 11:58:06,843 DEBUG [BpelServerImpl] handleWorkEvent: InvokeCheck event for mexid hqejbhcnphr3w17uvkwuxl DEBUG - GeronimoLog.debug(66) | Dangling invocation (mexId=hqejbhcnphr3w17uvkwuxl), forcing it into a failed state. 11:58:07,062 DEBUG [BpelServerImpl] Dangling invocation (mexId=hqejbhcnphr3w17uvkwuxl), forcing it into a failed state. DEBUG - GeronimoLog.debug(66) | replyWithFailure mex=hqejbhcnphr3w17uvkwuxl 11:58:07,062 DEBUG [PartnerRoleMessageExchangeImpl] replyWithFailure mex=hqejbhcnphr3w17uvkwuxl But the process was not resumed, while as per the flow , it shd retry in case of failure. Regards, Rosalin On Tue, Jan 6, 2009 at 11:34 PM, Alex Boisvert (JIRA) <[email protected]>wrote: > > [ > https://issues.apache.org/jira/browse/ODE-473?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12661233#action_12661233] > > Alex Boisvert commented on ODE-473: > ----------------------------------- > > On Tue, Jan 6, 2009 at 1:42 AM, Rosalin Pattnaik < > [email protected] > > <snip> > > > > Right, this is expected. > > > > <snip> > > The expected sequence of action should be the following, > 1) process invokes service > 2) server crashes before invoke is successful or timeouts > 3) server is restarted > 4) when the message exchange times-out (after 120 seconds, I believe) a > scheduled job should mark the exchange as failed and then proceed according > to your activity failure instruction > 5) in your case, the invoke should be retried automatically. > > Can you test with the latest Ode 1.x branch as requested? If you're not > getting this expected result then we'll look into it. > > alex > > > > If ode crashes while waiting for an response to an synchronous response > and while thr has yet not been any timeout , it does not resume on > restarting the ode. > > > -------------------------------------------------------------------------------------------------------------------------------------------------------------- > > > > Key: ODE-473 > > URL: https://issues.apache.org/jira/browse/ODE-473 > > Project: ODE > > Issue Type: Bug > > Components: BPEL Runtime > > Reporter: Rosalin Pattnaik > > Priority: Critical > > > > If we have bpel extension for retry on timeout with an invoke action , > but the ode crashes before the timeout has occured .And thus there is > neither a timeout exception nor has it received the response which might > have been lost in the time it was down .When the ode restarts which of the > following shd occur: > > * The ode resumes from the invoke state ,waits for the response , > timesout and retries, get the response and proceed normally. > > * The process is not resumed. > > In my case the process is not resumed in case timeout exception has not > occured , but in case the timeout has occured and then the ode crashes then > it retries gets the response and then completes succesfully. > > [N.B It resumed in case there was an entry into the ode_activity_recovery > table , or else it didnt] > > Please find the BPEL and external service used as attachment with the > query posted at ode user forum : > > > http://www.nabble.com/If-ode-crashes-while-waiting-for-an-response-to-an-synchronous-response-and-while-thr-has-yet-not-been-any-timeout-%2C-Does-it-on-restarting-resume-the-process-and-the-timesout-the-invoke-activity--td21248664.html > > -- > This message is automatically generated by JIRA. > - > You can reply to this email to add a comment to the issue online. > >
