[
https://issues.apache.org/jira/browse/HDDS-14205?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ivan Andika updated HDDS-14205:
-------------------------------
Description:
Unlike Ratis pipeline, Streaming Write Pipeline will fail the whole write if
any of the datanode cannot are not able to write the data. However, currently
we still use the same fallback MAJORITY_COMMITTED logic in Ratis pipeline. We
might consider to enforce ALL_COMMITTED for StreamingWritePipeline.
We might need send another ALL_COMMITTED or simply fail if
NotReplicatedException is received.
was:
Unlike Ratis pipeline, Streaming Write Pipeline will fail the whole write if
any of the datanode cannot are not able to write the data. However, currently
we still use the same fallback MAJORITY_COMMITTED logic in Ratis pipeline.
We might need send another ALL_COMMITTED or simply fail if
NotReplicatedException is received.
> Streaming write PutBlock should always use ALL_COMMITTED
> --------------------------------------------------------
>
> Key: HDDS-14205
> URL: https://issues.apache.org/jira/browse/HDDS-14205
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Ivan Andika
> Assignee: Ivan Andika
> Priority: Major
>
> Unlike Ratis pipeline, Streaming Write Pipeline will fail the whole write if
> any of the datanode cannot are not able to write the data. However, currently
> we still use the same fallback MAJORITY_COMMITTED logic in Ratis pipeline. We
> might consider to enforce ALL_COMMITTED for StreamingWritePipeline.
> We might need send another ALL_COMMITTED or simply fail if
> NotReplicatedException is received.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
