Uploaded image for project: 'JS - JobScheduler'
  1. JS - JobScheduler
  2. JS-1212

JS Cluster order does not work as expected if first node is skipped

    XMLWordPrintable

Details

    Description

      Jobchain with two nodes i.e. A and B , was configured with distributed orders.
      In normal operation if two orders are submitted one after another, distributed cluster divide the orders equally among the cluster members.

      But if first node is skipped and two orders are submitted one after another , both the orders will be executed by the cluster member which receives the orders.

      JobScheduler also throws an error

      (CppLogger.scala:22) - SCHEDULER-303 Problem with database: SCHEDULER-360 Error when accessing database table SCHEDULER_ORDERS [SCHEDULER-379 Order CLUSTER_PROCESS/job_chain2:15143 is occupied by JobScheduler member 'CLUST_DJSMP01/MP:4431'] 

      s::scheduler::order::Order_subsystem_impl::try_load_order_from_database]
JobSchedulerLog4JAppender (system.out):  INFO [main] (CppLogger.scala:24) - (Database) SCHEDULER-907  Opening database: jdbc -id=spooler -class=oracle.jdbc.driver.OracleDriver jdbc:oracle:thin:@:1521:DORCL01 -user=sos_scheduler -password=sos
JobSchedulerLog4JAppender (system.out):  INFO [main] (CppLogger.scala:24) - (Database) SCHEDULER-807  Using database product Oracle
JobSchedulerLog4JAppender (system.out):  INFO [main] (CppLogger.scala:24) - (Database) SCHEDULER-957  Closing database
JobSchedulerLog4JAppender (system.out): ERROR [main] (CppLogger.scala:22) - SCHEDULER-303  Problem with database: SCHEDULER-360  Error when accessing database table SCHEDULER_ORDERS [SCHEDULER-379  Order CLUSTER_PROCESS/job_chain2:15143 is occupied by JobScheduler member 'CLUST_DJSMP01/MP:4431'] [so
s::scheduler::order::Order_subsystem_impl::try_load_order_from_database]
JobSchedulerLog4JAppender (system.out):  INFO [main] (CppLogger.scala:24) - (Database) SCHEDULER-907  Opening database: jdbc -id=spooler -class=oracle.jdbc.driver.OracleDriver jdbc:oracle:thin:@:1521:DORCL01 -user=sos_scheduler -password=xyz

      Please see attached example for the

      Attachments

        Issue Links

          relates to

          Feature - A feature of the product, which has yet to be developed. JS-1227 An order executed on any cluster member should be suspendable

          • Major - Major loss of function.
          • Released

          Fix - JS-1122 In a distributed orders cluster skipped nodes are executed if they are the first node in a job chain.

          • Major - Major loss of function.
          • Released

          Fix - JS-1235 Old job-microscheduling (of v1.6) is active again

          • Major - Major loss of function.
          • Released
          will be tested by

          Test Case - JS-1237 (TestCase) JS Cluster order does not work as expected if first node is skipped

          • Major - Major loss of function.
          • Released

          Activity

            People

              mp Mahendra Patidar
              mp Mahendra Patidar
              Joacim Zschimmer Joacim Zschimmer
              Mahendra Patidar Mahendra Patidar
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: