[
https://issues.apache.org/jira/browse/HDFS-7439?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14340327#comment-14340327
]
Tsz Wo Nicholas Sze commented on HDFS-7439:
-------------------------------------------
Thanks for the new patch. Some minor comments:
- There is a typo in DFSClient.inferChecksumTypeByReading: "tring" should be
"trying".
- Remove "for OP_READ_BLOCK" from DataTransferProtoUtil.checkBlockOpStatus(..)
and add it back to the message in RemoteBlockReader2.checkSuccess(..).
- Need to remove the original if-statement from Dispatcher.receiveResponse(..).
> Add BlockOpResponseProto's message to DFSClient's exception message
> -------------------------------------------------------------------
>
> Key: HDFS-7439
> URL: https://issues.apache.org/jira/browse/HDFS-7439
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Ming Ma
> Assignee: Takanobu Asanuma
> Priority: Minor
> Attachments: HDFS-7439.1.patch, HDFS-7439.2.patch
>
>
> When (BlockOpResponseProto#getStatus() != SUCCESS), it helps with debugging
> if DFSClient can add BlockOpResponseProto's message to the exception message
> applications will get. For example, instead of
> {noformat}
> throw new IOException("Got error for OP_READ_BLOCK, self="
> + peer.getLocalAddressString() + ", remote="
> + peer.getRemoteAddressString() + ", for file " + file
> + ", for pool " + block.getBlockPoolId() + " block "
> + block.getBlockId() + "_" + block.getGenerationStamp());
> {noformat}
> It could be,
> {noformat}
> throw new IOException("Got error for OP_READ_BLOCK, self="
> + peer.getLocalAddressString() + ", remote="
> + peer.getRemoteAddressString() + ", for file " + file
> + ", for pool " + block.getBlockPoolId() + " block "
> + block.getBlockId() + "_" + block.getGenerationStamp()
> + ", status message " + status.getMessage());
> {noformat}
> We might want to check out all the references to BlockOpResponseProto in
> DFSClient.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)