[
https://issues.apache.org/jira/browse/HDFS-13042?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16337185#comment-16337185
]
Yiqun Lin commented on HDFS-13042:
----------------------------------
Hi [~elgoiri], the patch overall looks good. There are some minor comments:
*DFSConfigKeys.java*
line1285: Document this config in file {{hdfs-default.xml}}.
*Router.java*
line126: Can we define a new state named {{UNINITIALIZED}} to replace {{NONE}}
in {{RouterServiceState}}?
*RouterHeartbeatService.java*
line63: Can we rename method to {{updateStateAsync}}? That will be better for
understanding.
line143: Why not define a new setting for control the interval of updating
router state? There will be some users want to let records cache updating
slowly than router state. When introducing the new config, it will be good to
support time units suffixed as well.
*RouterState.java*
line56: Here we should update the modification time of the record. Otherwise,
the router state of record will always become expired state.
*TestStateStoreRouterState.java*
line131: Can you add the test case that the state changed from {{expired}} to
{{running}} after one successful router heartbeat? This should be our expected
behavior.
line182: Can you do the additionally router-address check for queried
RouterStates?
> RBF: Heartbeat Router State
> ---------------------------
>
> Key: HDFS-13042
> URL: https://issues.apache.org/jira/browse/HDFS-13042
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Íñigo Goiri
> Assignee: Íñigo Goiri
> Priority: Major
> Attachments: HDFS-13042.000.patch, HDFS-13042.001.patch
>
>
> The Router should heartbeat its state to the State Store.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
