View Issue Details

IDProjectCategoryView StatusLast Update
0003305SymmetricDSBugpublic2017-11-10 08:37
ReportermmichalekAssigned Tommichalek 
PrioritynormalSeverityminorReproducibilityhave not tried
Status assignedResolutionopen 
Product Version3.8.0 
Target VersionFixed in Version 
Summary0003305: Removing a channel can cause errors during push/pull
DescriptionGiven preconditions:
1) cluster.lock.enabled=true
2) Channels set up using queue names other than default.
3) Sync some data on a non-default queue so that rows get created in sym_node_communication.

Now, assign trigger back to the default channel, and remove the non-default channel. The rows pointing at the new queue will hang around in the sym_node_communication and cause errors like this during PUSH/PULL:

2017-11-06 12:47:39,579 ERROR [store-002] [NodeCommunicationService] [store-002-pull-product-5] Failed to execute PULL for node 000 and channel product
java.lang.NullPointerException
    at org.jumpmind.symmetric.service.impl.PullService.execute(PullService.java:117)
    at org.jumpmind.symmetric.service.impl.NodeCommunicationService$1.run(NodeCommunicationService.java:519)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
    at java.lang.Thread.run(Thread.java:745)
2017-11-06 12:47:39,582 ERROR [store-002] [NodeCommunicationService] [store-002-push-product-5] Failed to execute PUSH for node 000 and channel product
java.lang.NullPointerException
    at org.jumpmind.symmetric.service.impl.PushService.execute(PushService.java:145)
    at org.jumpmind.symmetric.service.impl.NodeCommunicationService$1.run(NodeCommunicationService.java:519)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
    at java.lang.Thread.run(Thread.java:745)
TagsNo tags attached.

Activities

ivan.konev

2017-11-10 07:14

reporter   ~0001078

Strongly agreed. Got same issue.

ivan.konev

2017-11-10 07:22

reporter   ~0001079

Sure that can be cleaned out manually by purging tables/correcting table with triggers or just disable cluster.lock.enable if not using clustering, anyway clustering gives only failover functionality.

ivan.konev

2017-11-10 08:37

reporter   ~0001080

Additional - does not required removing a channel. Just change queue name back or to something else.

Issue History

Date Modified Username Field Change
2017-11-06 12:57 mmichalek New Issue
2017-11-06 12:57 mmichalek Status new => assigned
2017-11-06 12:57 mmichalek Assigned To => mmichalek
2017-11-10 07:14 ivan.konev Note Added: 0001078
2017-11-10 07:22 ivan.konev Note Added: 0001079
2017-11-10 08:37 ivan.konev Note Added: 0001080