FSNamesystem.countNodes(Block b) should be synchronized.
--------------------------------------------------------
Key: HADOOP-4972
URL: https://issues.apache.org/jira/browse/HADOOP-4972
Project: Hadoop Core
Issue Type: Bug
Components: dfs
Reporter: Tsz Wo (Nicholas), SZE
(Copied a
[comment|https://issues.apache.org/jira/browse/HADOOP-4840?focusedCommentId=12657242#action_12657242]
from HADOOP-4840.)
FSNamesystem.countNodes(..) is called in many places including:
* FSNamesystem.addStoredBlock(Block, DatanodeDescriptor, DatanodeDescriptor)
* FSNamesystem.checkReplicationFactor(INodeFile)
* FSNamesystem.decrementSafeBlockCount(Block)
* FSNamesystem.getBlockLocationsInternal(String, INodeFile, long, long,
int, boolean)
* FSNamesystem.invalidateBlock(Block, DatanodeInfo)
* FSNamesystem.isReplicationInProgress(DatanodeDescriptor)
* FSNamesystem.markBlockAsCorrupt(Block, DatanodeInfo)
* FSNamesystem.processMisReplicatedBlocks()
* FSNamesystem.processPendingReplications()
* FSNamesystem.updateNeededReplications(Block, int, int)
However, some of them, e.g. getBlockLocationsInternal, call countNodes(..)
without owning the fsnamesystem lock before calling . It may causes NPE in
runtime.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
