[
https://issues.apache.org/jira/browse/AMBARI-17529?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15645669#comment-15645669
]
Hudson commented on AMBARI-17529:
---------------------------------
FAILURE: Integrated in Jenkins build Ambari-branch-2.5 #278 (See
[https://builds.apache.org/job/Ambari-branch-2.5/278/])
AMBARI-17529. Ambari LogSearch REST Layer should not queue up requests
(rnettleton:
[http://git-wip-us.apache.org/repos/asf?p=ambari.git&a=commit&h=d80e057b26e32eb6e942f7b972bdce69420d6028])
* (edit)
ambari-server/src/main/java/org/apache/ambari/server/controller/logging/LogSearchDataRetrievalService.java
* (edit)
ambari-server/src/test/java/org/apache/ambari/server/controller/logging/LogSearchDataRetrievalServiceTest.java
> Ambari LogSearch REST Layer should not queue up requests that have already
> been made
> ------------------------------------------------------------------------------------
>
> Key: AMBARI-17529
> URL: https://issues.apache.org/jira/browse/AMBARI-17529
> Project: Ambari
> Issue Type: Bug
> Components: ambari-server
> Affects Versions: 2.4.0
> Reporter: Robert Nettleton
> Assignee: Robert Nettleton
> Priority: Critical
> Fix For: 2.5.0
>
> Attachments: AMBARI-17529.patch.2
>
>
> The review for the fix for AMBARI-17510 brought up some interesting issues
> that should be investigated and addressed in a separate patch.
> From [~jonathan.hurley]'s comments on https://reviews.apache.org/r/49474/:
> "Based on what we've seen, the requests to the LOGSEARCH endpoint don't
> return until the connection times out (which could be never depending on
> what's configured). This will then cause this Executor to fill it's queue
> with a backlog of requests.
> Instead, maybe it would be better to:
> - Only enqueue if the request isn't already enqueued
> - Use a bounded executor and rejection policy"
> This work will involve investigating the best approach towards limiting the
> possibility of duplicate requests being made to the LogSearch REST service,
> either by making a check prior to adding the request, or by using a bounded
> executor in the way described above.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
