I followed exactly Sheetal's code and the problems still remain.
It seems a little strange behavior that I try to explain shortly:
A node (say node1) instantiates two cluster (group1 and group2). Another
node (node2) joins group1; so the two views are:
group1: [node1, node2]
group2: [node1]
node1 and node2 do nothing until node2 crashes without doing a clean
shutdown.
When node1 get notified of node2 failure this scenario happens:
26-feb-2008 21.19.57 com.sun.enterprise.ee.cms.impl.common.Router
notifyFailureSuspectedAction
INFO: Sending FailureSuspectedSignals to registered Actions. Member:node2...
26-feb-2008 21.20.00 com.sun.enterprise.ee.cms.impl.jxta.ViewWindow
getMemberTokens
INFO: GMS View Change Received: Members in view (before change analysis)
are :
26-feb-2008 21.20.00 com.sun.enterprise.ee.cms.impl.jxta.ViewWindow
newViewObserved
INFO: Analyzing new membership snapshot received as part of event :
FAILURE_EVENT
26-feb-2008 21.20.00 com.sun.enterprise.ee.cms.impl.jxta.ViewWindow
addFailureSignals
INFO: The following member has failed: node1
and the two views are:
group1: [node1, node2]
group2: []
For Shoal, node1 seems to have crashed even if it is actually up and
running. So the view of group2 is update eliminating node1. The view of
group1 is not updated at all, containing an element that isn't running
anymore.
In addition, node1 can still access group1 (i.e. to get member list and
so on)
Emanuele
Shreedhar Ganapathy wrote:
> Hi Emanuele
> Take at look at
> shoal/gms/tests/com/sun/enterprise/ee/cms/tests/multigroupjoin/MultiGroupJoinTest.java
> that Sheetal checked in recently.
>
> That should help.
> regards
> Shreedhar
>
> Emanuele wrote:
>> Hi,
>> In my app I need that a node instantiate three different group. Here
>> it is the code:
>>
>> Runnable gms1R, gms2R, gms3R = null;
>> Thread gms1T, gms2T, gms3T = null;
>> Properties properties = new Properties();
>> gms1R = GMSFactory.startGMSModule(
>> "node1",
>> "MainCluster",
>> GroupManagementService.MemberType.CORE,
>> properties);
>> gms1T = new Thread(gms1R, "GMS_Main_Cluster_Thread");
>> gms2R = GMSFactory.startGMSModule(
>> "node1",
>> "group1",
>> GroupManagementService.MemberType.CORE,
>> properties);
>> gms2T = new Thread(gms2R, "GMS_My_Cluster_Thread");
>> gms3R = GMSFactory.startGMSModule(
>> "node1",
>> "SupervisorCluster",
>> GroupManagementService.MemberType.CORE,
>> properties);
>> gms3T = new Thread(gms3R, "GMS_All_Supervisor_Cluster_Thread");
>> gms1T.start();
>> gms2T.start();
>> gms3T.start();
>>
>> It happens that the node always fails in two groups and I receive the
>> node notification failure.
>> It seems that the node remains in the two groups in which it has
>> failed even if
>> GroupManagementService.getGroupHandle().getAllCurrentMembers()
>> returns a zero-size list. Moreover
>> GMSFactory.getAllGMSInstancesForMember() returns all the three groups
>> instantiated.
>>
>> Is this that correct way to instantiate multiple cluster or I'm
>> missing something?
>> Thanks.
>>
>> Emanuele
>>
>>
>> --
>> Email.it, the professional e-mail, gratis per te: http://www.email.it/f
>>
>> Sponsor:
>> Analisi di Bilancio: come capire correttamente un bilancio aziendale
>> * Esempi, modelli pratici, software utili Clicca qui:
>> http://adv.email.it/cgi-bin/foclick.cgi?mid=7594&d=26-2
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_shoal.dev.java.net
>> For additional commands, e-mail: users-help_at_shoal.dev.java.net
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_shoal.dev.java.net
> For additional commands, e-mail: users-help_at_shoal.dev.java.net
>
>
--
Email.it, the professional e-mail, gratis per te:
http://www.email.it/f
Sponsor:
Vuoi imparare a scrivere un Business Plan vincente? Pratico Videcorso multimediale con modelli, esempi e consigli degli esperti
Clicca qui:
http://adv.email.it/cgi-bin/foclick.cgi?mid=7593&d=26-2