You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@chukwa.apache.org by "shreyas subramanya (JIRA)" <ji...@apache.org> on 2014/07/01 01:43:24 UTC
[jira] [Commented] (CHUKWA-714) Ability to make memory queue
configurable
[ https://issues.apache.org/jira/browse/CHUKWA-714?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14048316#comment-14048316 ]
shreyas subramanya commented on CHUKWA-714:
-------------------------------------------
With this patch, we have additional configuration parameters:
"chukwaAgent.chunk.queue" - the implementation of ChunkQueue to be used by the agent. Defaults to "org.apache.hadoop.chukwa.datacollection.agent.MemLimitQueue"
"chukwaAgent.chunk.queue.limit" - the size of chunk queue in bytes. Defaults to 10MB (10485760)
Any new implementation of ChunkQueue can have additional configuration properties, by implementing a constructor with org.apache.hadoop.conf.Configuration parameter.
> Ability to make memory queue configurable
> -------------------------------------------
>
> Key: CHUKWA-714
> URL: https://issues.apache.org/jira/browse/CHUKWA-714
> Project: Chukwa
> Issue Type: New Feature
> Components: Data Collection
> Affects Versions: 0.5.0, 0.6.0
> Reporter: shreyas subramanya
> Assignee: shreyas subramanya
> Priority: Minor
> Fix For: 0.6.0
>
> Attachments: CHUKWA-714.patch, CHUKWA-714.patch2
>
>
> The current implementation of the internal memory queue has blocking semantics for the sender. This has the potential of blocking the sender (such as log4j socket appender application) when the queue is full. This jira provides an implementation of a non blocking queue and exposes configuration parameters to switch the queue implementation as well as specify the max size of the queue.
--
This message was sent by Atlassian JIRA
(v6.2#6252)