[
https://issues.apache.org/jira/browse/IMPALA-10240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17214064#comment-17214064
]
Sahil Takiar commented on IMPALA-10240:
---------------------------------------
Would be nice to document what exactly happens when a node is remove from the
cluster membership - e.g. all queries running on that node are either cancelled
or retried (they are retried if transparent query retries are enabled). This
should cover the scenario where a coordinator fails as well (e.g. all queries
die, and the executors eventually timeout all fragments and cancel them).
> Impala Doc: Add docs for cluster membership statestore heartbeats
> -----------------------------------------------------------------
>
> Key: IMPALA-10240
> URL: https://issues.apache.org/jira/browse/IMPALA-10240
> Project: IMPALA
> Issue Type: Sub-task
> Reporter: Sahil Takiar
> Priority: Major
>
> I don't see many docs explaining how the current cluster membership logic
> works (e.g. via the statestored heartbeats). Would be nice to include a high
> level explanation along with how to configure the heartbeat threshold.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
