You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@kudu.apache.org by "Brian McDevitt (Jira)" <ji...@apache.org> on 2020/05/18 14:41:00 UTC
[jira] [Commented] (KUDU-3122) Add a flag to disable the
KuduContext accumulators
[ https://issues.apache.org/jira/browse/KUDU-3122?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17110351#comment-17110351 ]
Brian McDevitt commented on KUDU-3122:
--------------------------------------
[~granthenke] Does it make sense to have a single flag to enable/disable all the KuduContext accumulators? Or group the accumulators by type and have a flag for each?
row accumulator - Row
table operation counts (insert, upsert, update, delete) - Map[string, long]
histogram - HdrHistogram
> Add a flag to disable the KuduContext accumulators
> --------------------------------------------------
>
> Key: KUDU-3122
> URL: https://issues.apache.org/jira/browse/KUDU-3122
> Project: Kudu
> Issue Type: Improvement
> Components: java, spark
> Reporter: Brian McDevitt
> Assignee: Brian McDevitt
> Priority: Major
>
> Based on the discussion in [https://gerrit.cloudera.org/#/c/15882/] this enhancement would allow users to disable the Spark accumulators in high memory usage applications.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)