ProActive Scheduler

Sometimes a ClassCastException occur in the Resource Manager

Details

  • Description:
    Hide
    After submitting a Job_8_task, I get a ClassCastException in the RM console.
    It was not a critical exception as the RM still works normally after that.

    I was not able to reproduce the bug.


    Starting Resource Manager, Please wait...
    (Once started, press 'e' to shutdown)
    Resource Manager successfully created on pamr://3:-1/

    Exception in thread "Users pinger" java.lang.ClassCastException: org.objectweb.proactive.core.body.UniversalBodyRemoteObjectAdapter cannot be cast to java.lang.Comparable
            at java.util.TreeMap.put(TreeMap.java:542)
            at java.util.TreeSet.add(TreeSet.java:238)
            at java.util.AbstractCollection.addAll(AbstractCollection.java:305)
            at java.util.TreeSet.addAll(TreeSet.java:295)
            at org.ow2.proactive.resourcemanager.frontend.RMUserImpl$Pinger.run(RMUserImpl.java:347)
    Show
    After submitting a Job_8_task, I get a ClassCastException in the RM console. It was not a critical exception as the RM still works normally after that. I was not able to reproduce the bug. Starting Resource Manager, Please wait... (Once started, press 'e' to shutdown) Resource Manager successfully created on pamr://3:-1/ Exception in thread "Users pinger" java.lang.ClassCastException: org.objectweb.proactive.core.body.UniversalBodyRemoteObjectAdapter cannot be cast to java.lang.Comparable         at java.util.TreeMap.put(TreeMap.java:542)         at java.util.TreeSet.add(TreeSet.java:238)         at java.util.AbstractCollection.addAll(AbstractCollection.java:305)         at java.util.TreeSet.addAll(TreeSet.java:295)         at org.ow2.proactive.resourcemanager.frontend.RMUserImpl$Pinger.run(RMUserImpl.java:347)

Activity

Hide
Oleg Smirnov added a comment - 07/May/09 5:56 PM
seems that it's related to my last changes
Show
Oleg Smirnov added a comment - 07/May/09 5:56 PM seems that it's related to my last changes
Hide
Brian Amedro added a comment - 07/May/09 6:14 PM
Bug was occuring when running over the new PAMR protocol. Here is an explanation of how to reproduce the bug :
1- Start a PAMR Router
$ ./startRouter.sh
... [addr=/138.96.218.202,localport=33647]

2- Add theses properties on the start of the RM, Scheduler, and Node
proactive.communication.protocol=pamr
proactive.net.router.address=138.96.218.202
proactive.net.router.port=33647

3- Start a node
$ ./startNode.sh node1
 ... pamr://0/node1 ...

4 - Start the RM
$ ./startRM.sh
 .... pamr://3/ ....
$ ./adminRM.sh -u pamr://3/

> addnode("pamr://0/node1","Default")
> listnodes()


5- Start the Scheduler
$ ./startScheduler.sh -u pamr://3/
 ... pamr://6:1099/ ...
$ ./adminScheduler.sh -u pamr://6/

> submit("/home/bamedro/Experimentations/AmazonEC2/scheduling/samples/jobs_descriptors/Job_8_tasks.xml")
> result(1)



Bug will occur in the RM core console window
Show
Brian Amedro added a comment - 07/May/09 6:14 PM Bug was occuring when running over the new PAMR protocol. Here is an explanation of how to reproduce the bug : 1- Start a PAMR Router $ ./startRouter.sh ... [addr=/138.96.218.202,localport=33647] 2- Add theses properties on the start of the RM, Scheduler, and Node proactive.communication.protocol=pamr proactive.net.router.address=138.96.218.202 proactive.net.router.port=33647 3- Start a node $ ./startNode.sh node1  ... pamr://0/node1 ... 4 - Start the RM $ ./startRM.sh  .... pamr://3/ .... $ ./adminRM.sh -u pamr://3/ > addnode("pamr://0/node1","Default") > listnodes() 5- Start the Scheduler $ ./startScheduler.sh -u pamr://3/  ... pamr://6:1099/ ... $ ./adminScheduler.sh -u pamr://6/ > submit("/home/bamedro/Experimentations/AmazonEC2/scheduling/samples/jobs_descriptors/Job_8_tasks.xml") > result(1) Bug will occur in the RM core console window

People

Dates

  • Created:
    07/May/09 5:54 PM
    Updated:
    04/Nov/09 5:01 PM
    Resolved:
    07/May/09 6:34 PM