[
https://issues.apache.org/jira/browse/HADOOP-5412?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12679395#action_12679395
]
Hairong Kuang commented on HADOOP-5412:
---------------------------------------
The problem was that a simulated DataNode allows two concurrent writes to the
same block. In the above case, NN happened to schedule two concurrent
replications of a block to the same datanode (127.0.0.1:48095). The first
replication succeeded but the second one failed on the error: Finalizing a
block that has already been finalized, which caused the block to be deleted on
the datanode. Subsequent request to replicate the block on this datanode failed
because the block was deleted. As a result, the block remained to stay
under-replicated before the test timed out.
> TestInjectionForSimulatedStorage oaccasionally fails on timeout
> ---------------------------------------------------------------
>
> Key: HADOOP-5412
> URL: https://issues.apache.org/jira/browse/HADOOP-5412
> Project: Hadoop Core
> Issue Type: Bug
> Affects Versions: 0.18.3
> Reporter: Hairong Kuang
> Assignee: Hairong Kuang
> Fix For: 0.18.4
>
>
> Occasionally TestInjectionForSimulatedStorage falls into an infinite loop,
> waiting for a block to reach its replication factor. The log repeatedly
> prints the following message:
> dfs.TestInjectionForSimulatedStorage
> (TestInjectionForSimulatedStorage.java:waitForBlockReplication(89)) - Not
> enough replicas for 2th block blk_6302924909504458109_1001 yet. Expecting 4,
> got 2.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
