[
https://issues.apache.org/jira/browse/FLUME-1124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13255881#comment-13255881
]
[email protected] commented on FLUME-1124:
------------------------------------------------------
bq. On 2012-04-17 15:44:12, Brock Noland wrote:
bq. >
flume-ng-core/src/main/java/org/apache/flume/lifecycle/LifecycleSupervisor.java,
line 68
bq. > <https://reviews.apache.org/r/4753/diff/3/?file=102382#file102382line68>
bq. >
bq. > This isn't your code but it looks like there are a few problems here.
bq. >
bq. > 1) monitorService on MonitorRunnable is not set so an NPE (which is
not being logged is thrown:
bq. >
bq. > java.lang.NullPointerException
bq. > at
org.apache.flume.lifecycle.LifecycleSupervisor$MonitorRunnable.run(LifecycleSupervisor.java:255)
bq. > at
java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:441)
bq. > at
java.util.concurrent.FutureTask$Sync.innerRunAndReset(FutureTask.java:317)
bq. > at
java.util.concurrent.FutureTask.runAndReset(FutureTask.java:150)
bq. > at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$101(ScheduledThreadPoolExecutor.java:98)
bq. > at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.runPeriodic(ScheduledThreadPoolExecutor.java:180)
bq. > at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:204)
bq. > at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
bq. > at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
bq. > at java.lang.Thread.run(Thread.java:662)
bq. >
bq. > 2) The NPE is not being logged.
bq. > 3) If the NPE did not occur we'd have a bigger issue as
scheduleAtFixedRate is called here but the runnable reschedules itself.
bq. > 4) Calling first supervise() and then start() creates two threads
monitoring this service.
bq.
bq. Hari Shreedharan wrote:
bq. Hi Brock,
bq.
bq. None of the issues are caused by this change. I can look into these
and fix them, it seems like it is because of the way the start method is
written. Here is a brief explanation:
bq. 1 and 2) This is caused by the start() function being called without
supervise being called. The reason we have not been seeing this seems to be
that when start is called, the different components have not been added. So it
never goes into the loop.
bq. 3) Agree this is an issue, but because of the explanation to (1 and 2)
above.
bq. 4) Agreed, we should not be doing this. I will update the relevant
code.
bq.
bq.
bq.
bq. Brock Noland wrote:
bq. I did not mean to imply this was caused by your change, but if we are
"fixing" the class we might as well fix these issues as well.
Yep, updated the jira.
- Hari
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/4753/#review6983
-----------------------------------------------------------
On 2012-04-17 15:06:58, Hari Shreedharan wrote:
bq.
bq. -----------------------------------------------------------
bq. This is an automatically generated e-mail. To reply, visit:
bq. https://reviews.apache.org/r/4753/
bq. -----------------------------------------------------------
bq.
bq. (Updated 2012-04-17 15:06:58)
bq.
bq.
bq. Review request for Flume.
bq.
bq.
bq. Summary
bq. -------
bq.
bq. All components are started by the same executor, which is different from
the executor service that checks the status of the components.
bq.
bq.
bq. This addresses bug FLUME-1124.
bq. https://issues.apache.org/jira/browse/FLUME-1124
bq.
bq.
bq. Diffs
bq. -----
bq.
bq.
flume-ng-core/src/main/java/org/apache/flume/lifecycle/LifecycleSupervisor.java
a7407e6
bq.
bq. Diff: https://reviews.apache.org/r/4753/diff
bq.
bq.
bq. Testing
bq. -------
bq.
bq.
bq. Thanks,
bq.
bq. Hari
bq.
bq.
> Lifecycle supervisor can cause thread contention, sometimes causing
> components to not startup.
> ----------------------------------------------------------------------------------------------
>
> Key: FLUME-1124
> URL: https://issues.apache.org/jira/browse/FLUME-1124
> Project: Flume
> Issue Type: Bug
> Components: Sinks+Sources
> Affects Versions: v1.1.0
> Reporter: Hari Shreedharan
> Assignee: Hari Shreedharan
> Fix For: v1.2.0
>
> Attachments: FLUME-1124-1.patch
>
>
> Based on Brock's review of my patch, the following issues have been added to
> this jira:
> This isn't your code but it looks like there are a few problems here.
> 1) monitorService on MonitorRunnable is not set so an NPE (which is not being
> logged is thrown:
> java.lang.NullPointerException
> at
> org.apache.flume.lifecycle.LifecycleSupervisor$MonitorRunnable.run(LifecycleSupervisor.java:255)
> at
> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:441)
> at
> java.util.concurrent.FutureTask$Sync.innerRunAndReset(FutureTask.java:317)
> at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:150)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$101(ScheduledThreadPoolExecutor.java:98)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.runPeriodic(ScheduledThreadPoolExecutor.java:180)
> at
> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:204)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
> at java.lang.Thread.run(Thread.java:662)
> 2) The NPE is not being logged.
> 3) If the NPE did not occur we'd have a bigger issue as scheduleAtFixedRate
> is called here but the runnable reschedules itself.
> 4) Calling first supervise() and then start() creates two threads monitoring
> this service.
--
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
