You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@karaf.apache.org by "Alexey Knyazev (JIRA)" <ji...@apache.org> on 2019/05/15 20:40:00 UTC
[jira] [Commented] (KARAF-6267) Group Join issue in Cellar
[ https://issues.apache.org/jira/browse/KARAF-6267?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16840752#comment-16840752 ]
Alexey Knyazev commented on KARAF-6267:
---------------------------------------
I have the same issue, my bundle uses jms client and when I am installing my feature on new karaf with cellar the hazelcast bundle stopped and cellar is not working correctly.
{{I do reserch and found this problem with *'org.apache.geronimo.specs/geronimo-jta_1.1_spec/1.1.1'* bundle. }}
{{This bundle expoted package *'javax.transaction.xa;version=1.1'*. }}
{{My feature depends on *'transaction'* feature, this feature install bundle exported package *'javax.transaction.xa;version=1.2'*. The _hazelcast_ bundle import package *javax.transaction.xa*. And when my feature is installing hazelcast try to restart and wiring with a new bundle. After this hazelcast is down.}}
{{I found some fast solition: before install *'cellar'* feature, need at first install *'transaction'* feature.}}
I think need to change _hazelcast_ feature definition in cellar repository, delete <bundle>mvn:org.apache.geronimo.specs/geronimo-jta_1.1_spec/1.1.1</bundle> and add <feature>transaction</feature>
> Group Join issue in Cellar
> --------------------------
>
> Key: KARAF-6267
> URL: https://issues.apache.org/jira/browse/KARAF-6267
> Project: Karaf
> Issue Type: Bug
> Components: cellar
> Affects Versions: cellar-4.1.3
> Environment: Windows 10
>
> Karaf Version: 4.2.0
> Cellar Version: 4.1.3
> Reporter: Kushal Gautam
> Assignee: Jean-Baptiste Onofré
> Priority: Major
>
> I have two nodes(management and worker) and three groups(default, management
> and workers). Both these nodes are new instances of Karaf with cellar
> installed on them, along with camel and jdbc. But the bundles using these
> services are not installed yet. So, it's a fresh karaf installation.
> The first node joined successfully to the management node and quit the
> default group. I tried to allocate the second node to the workers node as
> shown in the dump below. But, I am strangely getting an error time and
> again.
> karaf@root()> cluster:group-list
> | Group | Members
> --+------------+----------------------
> | management | 192.168.99.1:5701
> x | default | 192.168.99.1:5702(x)
> karaf@root()> cluster:group-create workers
> karaf@root()> cluster:group-list
> | Group | Members
> --+------------+----------------------
> | management | 192.168.99.1:5701
> | workers |
> x | default | 192.168.99.1:5702(x)
> karaf@root()> cluster:group-join workers
> No result received within given timeout
> karaf@root()> log:tail
> 13:29:42.862 INFO [hz.cellar.InvocationMonitorThread] [192.168.99.1]:5702
> [cellar] [3.9.1] Invocations:1 timeouts:1 backup-timeouts:0
> 13:30:31.539 ERROR [pool-14-thread-26] Error while dispatching task
> java.lang.NullPointerException: null
> at java.util.HashSet.<init>(HashSet.java:118) ~[?:?]
> at
> org.apache.karaf.cellar.hazelcast.HazelcastGroupManager.registerGroup(HazelcastGroupManager.java:467)
> ~[67:org.apache.karaf.cellar.hazelcast:4.1.2]
> at
> org.apache.karaf.cellar.core.control.ManageGroupCommandHandler.joinGroup(ManageGroupCommandHandler.java:91)
> ~[65:org.apache.karaf.cellar.core:4.1.2]
> at
> org.apache.karaf.cellar.core.control.ManageGroupCommandHandler.execute(ManageGroupCommandHandler.java:41)
> ~[65:org.apache.karaf.cellar.core:4.1.2]
> at
> org.apache.karaf.cellar.core.control.ManageGroupCommandHandler.execute(ManageGroupCommandHandler.java:27)
> ~[65:org.apache.karaf.cellar.core:4.1.2]
> at
> org.apache.karaf.cellar.core.command.CommandHandler.handle(CommandHandler.java:40)
> ~[65:org.apache.karaf.cellar.core:4.1.2]
> at
> org.apache.karaf.cellar.core.command.CommandHandler.handle(CommandHandler.java:28)
> ~[65:org.apache.karaf.cellar.core:4.1.2]
> at
> org.apache.karaf.cellar.core.event.EventDispatchTask.run(EventDispatchTask.java:67)
> [65:org.apache.karaf.cellar.core:4.1.2]
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> [?:?]
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> [?:?]
> at java.lang.Thread.run(Thread.java:745) [?:?]
> karaf@root()>
> cluster:group-list
> | Group | Members
> --+------------+----------------------
> | management | 192.168.99.1:5701
> | workers |
> x | default | 192.168.99.1:5702(x)
> On the management node, I have set the producer and consumer state to OFF
> Apart from that, I have added "org.ops4j.pax.transx.tm.geronimo" to the
> excluded config list in "org.apache.karaf.cellar.node" to avoid transaction
> lock issue.
> It would be great if somebody could help me figure out the issue here. I
> walked through the source code, but I could not see how exactly is it
> getting a null pointer.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)