[
https://issues.apache.org/jira/browse/HDFS-7348?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yi Liu updated HDFS-7348:
-------------------------
Attachment: HDFS-7348.003.patch
Update the patch for Zhe's comments.
I also remove the hardcode of (dataBlkNum, parityBlkNum, cellSize), since they
are now available in ECSchema from BlockECRecoveryCommand.
The test passed in my local environment and current tests include recovering
one or more parity blocks and recovering 3 data blocks, we can update
{{recoverTargets}} and related tests in follow-on after the decode
issue(HADOOP-11847) fixed.
The comments are addressed in the patch and some replies:
{quote}
I guess somewhere in the code (not necessarily in this path) we should assert
all internal blocks indeed share the same checksum?
{quote}
You are right, we can assert them, we can do it in a {{else}}.
{quote}
In the main run loop, success and success4Target look a little confusing ...
{quote}
Right, I rename {{success4Target}} to {{targetsStatus}} as you suggested, I
keep others and not use BitSet, since we need array access for them, which is
more efficient and convenient for code.
Besides, I add {{targetInputStreams}}, they are used in the sasl negotiation
when connecting to other DNs, and we'd better to close it too, even thought we
had closed it's socket.
> Erasure Coding: striped block recovery
> --------------------------------------
>
> Key: HDFS-7348
> URL: https://issues.apache.org/jira/browse/HDFS-7348
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: datanode
> Reporter: Kai Zheng
> Assignee: Yi Liu
> Attachments: ECWorker.java, HDFS-7348.001.patch, HDFS-7348.002.patch,
> HDFS-7348.003.patch
>
>
> This JIRA is to recover one or more missed striped block in the striped block
> group.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
