[
https://issues.apache.org/jira/browse/HDFS-8838?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14648802#comment-14648802
]
Li Bo commented on HDFS-8838:
-----------------------------
hi, Nicholas
I think you can commit your patch first and I will update mine after that.
Some points :
1. {{DFSStripedOutputStream#getNumBlockWriteRetry}} returns 0, which
allows connecting to datanode only one time. I think we should allow the
connecting to be retied for several times. One way is to store the located
block getting from {{locateFollowingBlock()}}, and the following retries will
use the store one, no need to call {{locateFollowingBlock()}} again.
2. in {{TestDFSStripedOutputStreamWithFailure}}, you store the test length
in {{LENGTHS}}. But when I read the code, I have to calculate the length by
myself to see what kind the test is. So, how about adding some comments , or
directly show the file length in parameter such as {{testDatanodeFailure(4*
cellSize +123)}}?
> Tolerate datanode failures in DFSStripedOutputStream when the data length is
> small
> ----------------------------------------------------------------------------------
>
> Key: HDFS-8838
> URL: https://issues.apache.org/jira/browse/HDFS-8838
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: hdfs-client
> Reporter: Tsz Wo Nicholas Sze
> Assignee: Tsz Wo Nicholas Sze
> Attachments: h8838_20150729.patch
>
>
> Currently, DFSStripedOutputStream cannot tolerate datanode failures when the
> data length is small. We fix the bugs here and add more tests.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
