[NETCONF-196] Boron: Clustered setup sometimes fails to boot with "server is unhealthy" Created: 17/May/16  Updated: 15/Mar/19  Resolved: 01/Oct/18

Status: Resolved
Project: netconf
Component/s: restconf-nb
Affects Version/s: None
Fix Version/s: None

Type: Bug
Reporter: Jozef Behran Assignee: Jakub Morvay
Resolution: Won't Do Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All

Attachments: File 5896.tgz     File buggyrun.tgz    
External issue ID: 5896

 Description   

Sometimes the clustered setup fails to boot properly. Search in logs reveals numerous "server is unhealthy" messages. See attached file for an example.

 Comments   
Comment by Jozef Behran [ 17/May/16 ]

Attachment buggyrun.tgz has been added with description: Example of a failed cluster boot

Comment by Robert Varga [ 17/May/16 ]

I cannot see any such message in the attached logs. Can you provide the offending exception?

Comment by Jozef Behran [ 18/May/16 ]

It turns out the original "failed cluster boot" attachment is not from the "failed cluster boot" runs at all (this is why it is now flagged as "obsolete"). Attached is the console outputs from the 2 test runs where this bug happened (the console output is all that was left). What is suspect is a "NullPointerException: Null service reference is not allowed" happening.

Comment by Jozef Behran [ 18/May/16 ]

Attachment 5896.tgz has been added with description: The correct outputs from the 2 failed tests.

Comment by Tom Pantelis [ 15/Jun/16 ]

restconf is failing to push:

java.lang.NullPointerException: A null service reference is not allowed.
at org.eclipse.osgi.framework.internal.core.BundleContextImpl.getService(BundleContextImpl.java:586)
at org.opendaylight.restconf.rest.RestConnectorProvider.getObjectFromBundleContext(RestConnectorProvider.java:50)
at org.opendaylight.restconf.rest.RestConnectorProvider.onSessionInitiated(RestConnectorProvider.java:37)
at org.opendaylight.controller.sal.dom.broker.BrokerImpl.registerProvider(BrokerImpl.java:159)
at org.opendaylight.RestconfWrapperProviders.registerProviders(RestconfWrapperProviders.java:63)
at org.opendaylight.controller.config.yang.md.sal.rest.connector.RestConnectorModule.createInstance(RestConnectorModule.java:34)
at org.opendaylight.controller.config.spi.AbstractModule.getInstance(AbstractModule.java:82)

2016-05-17 11:08:49,554 | ERROR | config-pusher | ConfigRegistryImpl | 109 - org.opendaylight.controller.config-manager - 0.5.0.SNAPSHOT | Configuration Transaction failed on 2PC, server is unhealthy
java.lang.IllegalStateException: Error - getInstance() failed for ModuleIdentifier

{factoryName='rest-connector-impl', instanceName='rest-connector-default-impl'}

in transaction TransactionIdentifier

{name='ConfigTransaction-73-75'}

This doesn't look related to clustering. This bug should be moved to the netconf project.

Comment by Jakub Morvay [ 01/Oct/18 ]

From the attached logs and log snippet mentioned by Tom P. it looks like this is related to config subsystem. Config subsystem is not supported nor present anymore.

Generated at Wed Feb 07 20:14:24 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.