[ 
https://issues.apache.org/jira/browse/FLUME-1136?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13257295#comment-13257295
 ] 

Hari Shreedharan commented on FLUME-1136:
-----------------------------------------

Before patch, logs on reconfiguration:
2012-04-18 23:12:38,515 INFO lifecycle.LifecycleSupervisor: Component has 
already been stopped SinkRunner: { 
policy:org.apache.flume.sink.DefaultSinkProcessor@311671b2 counterGroup:{ 
name:null counters:{runner.backoffs.consecutive=6, runner.backoffs=10, 
runner.interruptions=1} } }
2012-04-18 23:12:38,860 INFO lifecycle.LifecycleSupervisor: Component has 
already been stopped EventDrivenSourceRunner: { source:AvroSource: { 
bindAddress:localhost port:25001 } }
2012-04-18 23:12:41,516 INFO lifecycle.LifecycleSupervisor: Component has 
already been stopped SinkRunner: { 
policy:org.apache.flume.sink.DefaultSinkProcessor@311671b2 counterGroup:{ 
name:null counters:{runner.backoffs.consecutive=6, runner.backoffs=10, 
runner.interruptions=1} } }
2012-04-18 23:12:41,861 INFO lifecycle.LifecycleSupervisor: Component has 
already been stopped EventDrivenSourceRunner: { source:AvroSource: { 
bindAddress:localhost port:25001 } }
2012-04-18 23:12:44,517 INFO lifecycle.LifecycleSupervisor: Component has 
already been stopped SinkRunner: { 
policy:org.apache.flume.sink.DefaultSinkProcessor@311671b2 counterGroup:{ 
name:null counters:{runner.backoffs.consecutive=6, runner.backoffs=10, 
runner.interruptions=1} } }
2012-04-18 23:12:44,863 INFO lifecycle.LifecycleSupervisor: Component has 
already been stopped EventDrivenSourceRunner: { source:AvroSource: { 
bindAddress:localhost port:25001 } }
2012-04-18 23:12:47,519 INFO lifecycle.LifecycleSupervisor: Component has 
already been stopped SinkRunner: { 
policy:org.apache.flume.sink.DefaultSinkProcessor@311671b2 counterGroup:{ 
name:null counters:{runner.backoffs.consecutive=6, runner.backoffs=10, 
runner.interruptions=1} } }


After patch:

2012-04-18 23:19:36,711 INFO sink.NullSink: Null sink stopping
2012-04-18 23:19:36,711 INFO sink.NullSink: Null sink stopped. Event metrics:{ 
name:null counters:{events.successful=25} }
2012-04-18 23:19:36,711 INFO lifecycle.LifecycleSupervisor: Stopping component: 
EventDrivenSourceRunner: { source:AvroSource: { bindAddress:localhost 
port:25001 } }
2012-04-18 23:19:36,711 INFO source.AvroSource: Avro source 
stopping:AvroSource: { bindAddress:localhost port:25001 }
2012-04-18 23:19:36,716 INFO source.AvroSource: Avro source 
starting:AvroSource: { bindAddress:localhost port:25001 }
2012-04-18 23:20:14,907 INFO ipc.NettyServer: [id: 0x30d1e7c2, /127.0.0.1:50120 
=> /127.0.0.1:25001] OPEN
2012-04-18 23:20:14,908 INFO ipc.NettyServer: [id: 0x30d1e7c2, /127.0.0.1:50120 
=> /127.0.0.1:25001] BOUND: /127.0.0.1:25001
2012-04-18 23:20:14,908 INFO ipc.NettyServer: [id: 0x30d1e7c2, /127.0.0.1:50120 
=> /127.0.0.1:25001] CONNECTED: /127.0.0.1:50120
2012-04-18 23:20:15,192 INFO sink.LoggerSink: Event: { headers:{} body: }


--
As you can see, after applying the above patch, there are no "Component has 
already been stopped" log messages, which were appearing every 3 seconds or so, 
for each component - for almost 40 seconds.
                
> Remove from executor service does not always remove the runnables from the 
> queue
> --------------------------------------------------------------------------------
>
>                 Key: FLUME-1136
>                 URL: https://issues.apache.org/jira/browse/FLUME-1136
>             Project: Flume
>          Issue Type: Bug
>          Components: Node
>    Affects Versions: v1.1.0
>            Reporter: Hari Shreedharan
>            Assignee: Hari Shreedharan
>            Priority: Minor
>             Fix For: v1.2.0
>
>         Attachments: FLUME-1136.patch
>
>
> It seems like ScheduledThreadPoolExecutor's remove() function sometimes does 
> not remove the runnable specified. This does not cause any functionality 
> problems, just spills lot of stuff into logs during reconfiguration, and can 
> cause runnables to be executed unnecessarily, even though they return 
> immediately.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to