[
https://issues.apache.org/jira/browse/HDFS-13062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16343896#comment-16343896
]
Hanisha Koneru commented on HDFS-13062:
---------------------------------------
{quote}Maintaining all directories in the localDir is as when JournalNode is
started, it will not be knowing which nameservice this journal node belongs to,
until we get some request from namenode like format eg., That is the reason for
maintaining all directories in localDir in Federated setup.
{quote}
Yes, but once the journals are set up and we have the information about which
dirs belong to the current JN, we should update the localDirs. I would suggest
maintaining two lists - one with all the possible journal dirs from the config
file and one with the actual dirs corresponding to the journal.
{quote}Regarding JournalStatus, as we are not creating the directory for the
journal, if it is not corresponding to this journalnode, so even if we iterate
through all localDir's, it will not display journalStatus of other JN.
{quote}
Even though the JN will not create the dir (in the example above, JN is {{jn1}}
and the dir is {{/disk2/jn}}), the dir might have been created manually or by
some other process. And in that case, its status will be wrongly displayed in
the jmx.
> Provide support for JN to use separate journal disk per namespace
> -----------------------------------------------------------------
>
> Key: HDFS-13062
> URL: https://issues.apache.org/jira/browse/HDFS-13062
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Bharat Viswanadham
> Assignee: Bharat Viswanadham
> Priority: Major
> Attachments: HDFS-13062.00.patch, HDFS-13062.01.patch,
> HDFS-13062.02.patch
>
>
> In Federated HA setup, provide support for separate journal disk for each
> namespace.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
