[
https://issues.apache.org/jira/browse/SPARK-3129?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14138032#comment-14138032
]
Hari Shreedharan commented on SPARK-3129:
-----------------------------------------
Thanks Matei for the background. I had considered some of the factors (like
executors always talking to the latest ones) - but I was not aware of the
distinct RDD ids etc.
TD and I discussed this offline and we agreed that the WAL would probably be
the best way to go. I am planning to do some benchmarking of appending data to
a 5-node HDFS cluster on EC2 today. Considering that HBase does use a WAL on
HDFS, my expectation is that the perf should be reasonable.
I will post the application on github and post a link here. I will run the
application and see how it goes. I will also post it here.
> Prevent data loss in Spark Streaming
> ------------------------------------
>
> Key: SPARK-3129
> URL: https://issues.apache.org/jira/browse/SPARK-3129
> Project: Spark
> Issue Type: New Feature
> Components: Streaming
> Reporter: Hari Shreedharan
> Assignee: Hari Shreedharan
> Attachments: SecurityFix.diff, StreamingPreventDataLoss.pdf
>
>
> Spark Streaming can small amounts of data when the driver goes down - and the
> sending system cannot re-send the data (or the data has already expired on
> the sender side). The document attached has more details.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
