[ 
https://issues.apache.org/jira/browse/YARN-9940?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16969831#comment-16969831
 ] 

kailiu_dev edited comment on YARN-9940 at 11/8/19 4:55 AM:
-----------------------------------------------------------

[~wilfreds] ,  you say right that  hadoop 2.7.2  has been 
{color:#172b4d}EOL'ed{color}, but can you please help me review my  my patch , 
check if can the code can {color:#172b4d}solving the  issure, beause my company 
need hadoop 2.7.2。{color}

*{color:#172b4d}in  my code:{color}*

for example , befor my change in hadoop 2.7.2:

*sort node is:*

synchronized (this)

{         Collections.sort(nodeIdList, nodeAvailableResourceComparator);     }

 

*but {color:#172b4d}completedContainer{color} is:*

writeLock.lock();

.....

node.releaseContainer(container); 

writeLock.unlock();

 

*after change my code:*

{color:#172b4d}completedContainer{color} is:
 
 writeLock.lock();

{color:#172b4d}if (continuousSchedulingEnabled) {
           synchronized (this) \{          node.releaseContainer(container);    
        }
         }{color}

{color:#172b4d}writeLock.unlock();{color}

 

{color:#ff0000}so  when releaseContainer , it should wait for geting schdule 
lock  when sorting node;{color}

{color:#ff0000}this  is my simple example,  the total code is in my 
patch.{color}

 

{color:#172b4d}so above on  you  say "the method 
{color}{{FairScheduler.completedContainer()}}{color:#172b4d} method is already 
synchronised adding a synchronised block inside that will not help{color}", I  
can  think that your meaning is that it is not helpful in hadoop2.9 beause of 
it has al used   {color:#172b4d}writeLock/readLock,{color}   but  it is helpful 
in hadoop2.7.2  ?
 
  
 
  

 

 


was (Author: kailiu_dev):
[~wilfreds] ,  you say right that  hadoop 2.7.2  has been 
{color:#172b4d}EOL'ed{color}, but can you please help me review my  my patch , 
check if can the code can {color:#172b4d}solving the  issure, beause my company 
need hadoop 2.7.2。{color}

*{color:#172b4d}in  my code:{color}*

for example , befor my change in hadoop 2.7.2:

*sort node is:*

synchronized (this) {
        Collections.sort(nodeIdList, nodeAvailableResourceComparator);
    }

but {color:#172b4d}completedContainer{color} is:

writeLock.lock();

.....

node.releaseContainer(container); 

writeLock.unlock();

 

*after change my code:*

{color:#172b4d}{color:#172b4d}completedContainer{color} is:{color}

{color:#172b4d}writeLock.lock();{color}

{color:#172b4d}if (continuousSchedulingEnabled) {
          synchronized (this) {
         node.releaseContainer(container); 
          }
        }{color}

{color:#172b4d}writeLock.unlock();{color}


 

{color:#FF0000}so  when releaseContainer , it should wait for geting schdule 
lock  when sorting node;{color}

{color:#FF0000}this  is my simple example,  the total code is in my 
patch.{color}

 

{color:#172b4d}so above on  you  say "t{color:#172b4d}he method 
{color}{{FairScheduler.completedContainer()}}{color:#172b4d} method is already 
synchronised adding a synchronised block inside that will not help{color}", I  
can  think that your meaning is that it is not helpful in hadoop2.9 beause of 
it has al used   {color:#172b4d}writeLock/readLock,{color}   but  it is helpful 
in hadoop2.7.2  ?{color}

 

{color:#172b4d} {color}

 

 

> avoid continuous scheduling thread crashes while sorting nodes get 
> 'Comparison method violates its general contract'
> --------------------------------------------------------------------------------------------------------------------
>
>                 Key: YARN-9940
>                 URL: https://issues.apache.org/jira/browse/YARN-9940
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: fairscheduler
>    Affects Versions: 2.7.2
>            Reporter: kailiu_dev
>            Assignee: kailiu_dev
>            Priority: Major
>         Attachments: YARN-9940-branch-2.7.2.001.patch
>
>
> 2019-10-16 09:14:51,215 ERROR 
> org.apache.hadoop.yarn.YarnUncaughtExceptionHandler: Thread 
> Thread[FairSchedulerContinuousScheduling,5,main] threw an Exception.
> java.lang.IllegalArgumentException: Comparison method violates its general 
> contract!
>         at java.util.TimSort.mergeHi(TimSort.java:868)
>         at java.util.TimSort.mergeAt(TimSort.java:485)
>         at java.util.TimSort.mergeForceCollapse(TimSort.java:426)
>         at java.util.TimSort.sort(TimSort.java:223)
>         at java.util.TimSort.sort(TimSort.java:173)
>         at java.util.Arrays.sort(Arrays.java:659)
>         at java.util.Collections.sort(Collections.java:217)
>         at 
> org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.FairScheduler.continuousSchedulingAttempt(FairScheduler.java:1117)
>         at 
> org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.FairScheduler$ContinuousSchedulingThread.run(FairScheduler.java:296)



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org

Reply via email to