[
https://issues.apache.org/jira/browse/HDFS-11493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16061514#comment-16061514
]
Xiaoyu Yao commented on HDFS-11493:
-----------------------------------
Thanks [~anu] for updating the patch. The patch looks pretty good to me. Here
are my comments:
*HadoopExecutors.java*
Line 104-133
Should we reuse the existing one in
ShutdownThreadsHelper.shutdownExecutorService?
*ScmConfigKeys.java*
Line 177: NIT: Dont -> Don’t
Line 186: Can we add some document on the other two configuration keys
introduced?
*OzoneConfigKeys.java*
NIT: Line 102-105 empty change
*ContainerReplicationManager.java*
Line 109: maxContainerReportThreads can be a local variable.
Line 133: NIT: javadoc incomplete
Line 191-199: this can be pulled into a generic CollectionUtil if it does not
exist in hadoop-common/hadoop-hdfs
Line 201: conf parameter can be removed as it is not used
Line 216: let’s change the log level to debug to avoid flooding the scm log.
Line 244-249: could this cause infinite loop when hitting unexpected
exceptions. Can we have a counter to limit the
number of time we will retry starting poolProcessThread? Also, adding a counter
as you mentioned in the TODO is a good
idea? Can we open a follow up JIRA on that?
Line 289: NIT: we can remove some of the javadocs on streams as they don’t
apply here.
*InProgressPools.java*
Should we rename it to InProgressPool if this is for a single pool being
processed?
Line 116: should we change “<“ to “>” to indicate that we have done waiting for
the maxWaitTime?
Line 207-228: one UNKNOWN node in the pool can cost 100s. Should we reduce the
maxTry from 1000 to 100 here?
*SCMNodeManager.java*
Line 405: missing /Unknown
*ReplicationDatanodeStateManager.java*
Line 78-81: do we miss the size/keycount for the ContainerInfo of the
ContainerReport?
> Ozone: SCM: Add the ability to handle container reports
> ---------------------------------------------------------
>
> Key: HDFS-11493
> URL: https://issues.apache.org/jira/browse/HDFS-11493
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ozone
> Affects Versions: HDFS-7240
> Reporter: Anu Engineer
> Assignee: Anu Engineer
> Attachments: container-replication-storage.pdf,
> exploring-scalability-scm.pdf, HDFS-11493-HDFS-7240.001.patch,
> HDFS-11493-HDFS-7240.002.patch, HDFS-11493-HDFS-7240.003.patch,
> HDFS-11493-HDFS-7240.004.patch
>
>
> Once a datanode sends the container report it is SCM's responsibility to
> determine if the replication levels are acceptable. If it is not, SCM should
> initiate a replication request to another datanode. This JIRA tracks how SCM
> handles a container report.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
