hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Íñigo Goiri (JIRA) <j...@apache.org>
Subject [jira] [Commented] (HDFS-14226) RBF: Setting attributes should set on all subclusters' directories.
Date Mon, 11 Feb 2019 18:05:00 GMT

    [ https://issues.apache.org/jira/browse/HDFS-14226?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16765225#comment-16765225

Íñigo Goiri commented on HDFS-14226:

Thanks for  [^HDFS-14226-HDFS-13891-02.patch]:
* Bad indentation on {{RouterClientProtocol#1720}}.
* We should fill the javadoc for {{RouterRpcServer#isMountEntry()}}.We should explain the
purpose too.
* What about a function (e.g., {{isInvokeConcurrent()}} or {{isInvokeAll()}}) where we explain
the whole issue which does the equivalent of {{return isPathAll() || isMountEntry();}}
* {{RouterRpcServer#isMountEntry()}} could also extract {{MountTableResolver mountTable =
(MountTableResolver) subclusterResolver;}} and make the if complete or even {{return mountPoints
!= null;}}.
* Add a javadoc to {{TestRouterRPCMultipleDestinationMountTableResolver#setUpTestEnviornment()}}
and create a nested folder with a nested file; this would cover the HASH_ALL case.
* Add a javadoc to {{testInvocation()}} explaining the purpose of {{boolean dirAll}}.
* {{TestRouterRPCMultipleDestinationMountTableResolver#159-179}} is pretty much repeated twice;
can we refactor to use {{nnFs0}} and {{nnFs1}} as a parameter? Similar for 186-199 and others
in this method.
* Add a comment to this block in TestRouterRPCMultipleDestinationMountTableResolver an ideally
TestRouterRPCMultipleDestinationMountTableResolver#180 should have a message on the {{assertTrue()}}.
* We are right now checking in the NN filesystem; we may want to first check in the routerFS

> RBF: Setting attributes should set on all subclusters' directories.
> -------------------------------------------------------------------
>                 Key: HDFS-14226
>                 URL: https://issues.apache.org/jira/browse/HDFS-14226
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>            Reporter: Takanobu Asanuma
>            Priority: Major
>              Labels: RBF
>         Attachments: HDFS-14226-HDFS-13891-01.patch, HDFS-14226-HDFS-13891-02.patch,
> Only one subcluster is set now.
> {noformat}
> // create a mount point of multiple subclusters
> hdfs dfsrouteradmin -add /all_data ns1 /data1
> hdfs dfsrouteradmin -add /all_data ns2 /data2
> hdfs ec -Dfs.defaultFS=hdfs://router:8888 -setPolicy -path /all_data -policy RS-3-2-1024k
> Set RS-3-2-1024k erasure coding policy on /all_data
> hdfs ec -Dfs.defaultFS=hdfs://router:8888 -getPolicy -path /all_data
> RS-3-2-1024k
> hdfs ec -Dfs.defaultFS=hdfs://ns1-namenode:8020 -getPolicy -path /data1
> RS-3-2-1024k
> hdfs ec -Dfs.defaultFS=hdfs://ns2-namenode:8020 -getPolicy -path /data2
> The erasure coding policy of /data2 is unspecified
> {noformat}

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org

View raw message