liaoyuxiangqin created HDFS-13479:
-------------------------------------
Summary: Simplify find StorageInfo logical operation in
BlocksMap::replaceBlock()
Key: HDFS-13479
URL: https://issues.apache.org/jira/browse/HDFS-13479
Project: Hadoop HDFS
Issue Type: Improvement
Components: namenode
Reporter: liaoyuxiangqin
When i read the replaceBlock() of BlocksMap class in hdfs-blockmanger, i found
the following find storage code could be more simplify and easy to understand.
{code:java|title=DataStreamer.java|borderStyle=solid}
for (int i = currentBlock.numNodes() - 1; i >= 0; i--) {
final DatanodeDescriptor dn = currentBlock.getDatanode(i);
final DatanodeStorageInfo storage = currentBlock.findStorageInfo(dn);
final boolean removed = storage.removeBlock(currentBlock);
Preconditions.checkState(removed, "currentBlock not found.");
final AddBlockResult result = storage.addBlock(newBlock);
Preconditions.checkState(result == AddBlockResult.ADDED,
"newBlock already exists.");
}
{code}
as described above code segmet, i find that need't get dn and we can find
storage by index directly, so i think this code logical could simplify more.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-dev-h...@hadoop.apache.org
