chukwa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ari Rabkin (JIRA)" <j...@apache.org>
Subject [jira] Commented: (CHUKWA-369) proposed reliability mechanism
Date Wed, 19 Aug 2009 02:05:14 GMT

    [ https://issues.apache.org/jira/browse/CHUKWA-369?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12744841#action_12744841
] 

Ari Rabkin commented on CHUKWA-369:
-----------------------------------

When you say "server a" and "sever b", do you mean Agent A and B, or Collector A and B?

For agent side first:  The collector knows which chunk is written where in the file, and it
tells each agent where in the file its data will be.

For the collector side: There's exactly one collector per sink file, and when an agent does
a post, the collector tells it the filename to track.

Alright.  I take your point that it's a mistake to have every collector doing an LS. I can
redesign so that only a handful of collectors are scanning the filesystem -- reducing the
load to an acceptable level. Would that address your concerns?

> proposed reliability mechanism
> ------------------------------
>
>                 Key: CHUKWA-369
>                 URL: https://issues.apache.org/jira/browse/CHUKWA-369
>             Project: Hadoop Chukwa
>          Issue Type: New Feature
>          Components: data collection
>    Affects Versions: 0.3.0
>            Reporter: Ari Rabkin
>            Assignee: Ari Rabkin
>             Fix For: 0.3.0
>
>         Attachments: delayedAcks.patch
>
>
> We like to say that Chukwa is a system for reliable log collection. It isn't, quite,
since we don't handle collector crashes.  Here's a proposed reliability mechanism.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message