Sandor Molnar created KNOX-2869:
-----------------------------------
Summary: Possible NPE at CM cluster configuration monitor startup
Key: KNOX-2869
URL: https://issues.apache.org/jira/browse/KNOX-2869
Project: Apache Knox
Issue Type: Bug
Affects Versions: 2.0.0
Reporter: Sandor Molnar
Assignee: Sandor Molnar
Fix For: 2.1.0
In some very rare cases, it could happen that CM service discovery-related
configuration file(s) were wrongly serialized on the file system and the next
time Knox starts the service initialization fails like this:
{noformat}
2023-01-30 10:39:01,733 FATAL knox.gateway (GatewayServer.java:main(193)) -
Failed to start gateway: java.lang.NullPointerException
java.lang.NullPointerException
at
java.base/java.util.concurrent.ConcurrentHashMap.computeIfAbsent(ConcurrentHashMap.java:1690)
at
org.apache.knox.gateway.topology.discovery.cm.monitor.ClusterConfigurationCache.addDiscoveryConfig(ClusterConfigurationCache.java:73)
at
org.apache.knox.gateway.topology.discovery.cm.monitor.ClouderaManagerClusterConfigurationMonitor.loadDiscoveryConfiguration(ClouderaManagerClusterConfigurationMonitor.java:186)
at
org.apache.knox.gateway.topology.discovery.cm.monitor.ClouderaManagerClusterConfigurationMonitor.<init>(ClouderaManagerClusterConfigurationMonitor.java:100)
at
org.apache.knox.gateway.topology.discovery.cm.monitor.ClouderaManagerClusterConfigurationMonitorProvider.newInstance(ClouderaManagerClusterConfigurationMonitorProvider.java:35)
at
org.apache.knox.gateway.services.topology.impl.DefaultClusterConfigurationMonitorService.init(DefaultClusterConfigurationMonitorService.java:44)
at
org.apache.knox.gateway.services.DefaultGatewayServices.init(DefaultGatewayServices.java:137)
at org.apache.knox.gateway.GatewayServer.main(GatewayServer.java:184)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.base/java.lang.reflect.Method.invoke(Method.java:566)
at
org.apache.knox.gateway.launcher.Invoker.invokeMainMethod(Invoker.java:68)
at org.apache.knox.gateway.launcher.Invoker.invoke(Invoker.java:39)
at org.apache.knox.gateway.launcher.Command.run(Command.java:99)
at org.apache.knox.gateway.launcher.Launcher.run(Launcher.java:75)
at
org.apache.knox.gateway.launcher.Launcher.main(Launcher.java:52){noformat}
In this particular case, the
{{$KNOX_DATA_DIR/cm-clusters/hCM_HOST_7183-Cluster_1.conf}} file was empty so
when Knox wanted to load any previously-persisted discovery configuration data
into its own in-memory cache, an NPE was thrown.
Although the chance for this to happen is quite low, Knox should handle this
situation better.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
