[ 
https://issues.apache.org/jira/browse/HDFS-17297?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Haiyang Hu updated HDFS-17297:
------------------------------
    Description: 
When call internalReleaseLease method:

{code:java}
boolean internalReleaseLease(
...
int minLocationsNum = 1;
if (lastBlock.isStriped()) {
  minLocationsNum = ((BlockInfoStriped) lastBlock).getRealDataBlockNum();
}

if (uc.getNumExpectedLocations() < minLocationsNum &&
    lastBlock.getNumBytes() == 0) {
  // There is no datanode reported to this block.
  // may be client have crashed before writing data to pipeline.
  // This blocks doesn't need any recovery.
  // We can remove this block and close the file.
  pendingFile.removeLastBlock(lastBlock);
  finalizeINodeFileUnderConstruction(src, pendingFile,
      iip.getLatestSnapshotId(), false); 
...
}
{code}
 if the condition `uc.getNumExpectedLocations() < minLocationsNum && 
lastBlock.getNumBytes() == 0` is met during the execution of UNDER_RECOVERY 
logic, the block is removed from the block list in the inode file and marked as 
deleted. 
However it is not removed from the BlocksMap, it may cause memory leak.

Therefore it is necessary to remove the block from the BlocksMap at this point 
as well.



  was:
When call internalReleaseLease method:

{code:java}
boolean internalReleaseLease(
...
int minLocationsNum = 1;
if (lastBlock.isStriped()) {
  minLocationsNum = ((BlockInfoStriped) lastBlock).getRealDataBlockNum();
}

if (uc.getNumExpectedLocations() < minLocationsNum &&
    lastBlock.getNumBytes() == 0) {
  // There is no datanode reported to this block.
  // may be client have crashed before writing data to pipeline.
  // This blocks doesn't need any recovery.
  // We can remove this block and close the file.
  pendingFile.removeLastBlock(lastBlock);
  finalizeINodeFileUnderConstruction(src, pendingFile,
      iip.getLatestSnapshotId(), false); 
...
}
{code}
 if the condition `uc.getNumExpectedLocations() < minLocationsNum && 
lastBlock.getNumBytes() == 0` is met during the execution of UNDER_RECOVERY 
logic, the block is removed from the block list in the inode file and marked as 
deleted. 
However it is not removed from the BlocksMap, it may cause memory leak.
Therefore it is necessary to remove the block from the BlocksMap at this point 
as well.




> The NameNode should remove block from the BlocksMap if the block is marked as 
> deleted.
> --------------------------------------------------------------------------------------
>
>                 Key: HDFS-17297
>                 URL: https://issues.apache.org/jira/browse/HDFS-17297
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>            Reporter: Haiyang Hu
>            Assignee: Haiyang Hu
>            Priority: Major
>
> When call internalReleaseLease method:
> {code:java}
> boolean internalReleaseLease(
> ...
> int minLocationsNum = 1;
> if (lastBlock.isStriped()) {
>   minLocationsNum = ((BlockInfoStriped) lastBlock).getRealDataBlockNum();
> }
> if (uc.getNumExpectedLocations() < minLocationsNum &&
>     lastBlock.getNumBytes() == 0) {
>   // There is no datanode reported to this block.
>   // may be client have crashed before writing data to pipeline.
>   // This blocks doesn't need any recovery.
>   // We can remove this block and close the file.
>   pendingFile.removeLastBlock(lastBlock);
>   finalizeINodeFileUnderConstruction(src, pendingFile,
>       iip.getLatestSnapshotId(), false); 
> ...
> }
> {code}
>  if the condition `uc.getNumExpectedLocations() < minLocationsNum && 
> lastBlock.getNumBytes() == 0` is met during the execution of UNDER_RECOVERY 
> logic, the block is removed from the block list in the inode file and marked 
> as deleted. 
> However it is not removed from the BlocksMap, it may cause memory leak.
> Therefore it is necessary to remove the block from the BlocksMap at this 
> point as well.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

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

Reply via email to