[
https://issues.apache.org/jira/browse/CHUKWA-415?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12781271#action_12781271
]
Ari Rabkin commented on CHUKWA-415:
-----------------------------------
I will commit this, barring objection.
> agent.sh should load adaptors from both checkpoint file and
> conf/initial_adaptors
> ---------------------------------------------------------------------------------
>
> Key: CHUKWA-415
> URL: https://issues.apache.org/jira/browse/CHUKWA-415
> Project: Hadoop Chukwa
> Issue Type: Improvement
> Components: data collection
> Reporter: Bill Graham
> Assignee: Ari Rabkin
> Fix For: 0.4.0
>
> Attachments: CHUKWA-415.patch
>
>
> Currently when agent.sh is run, the agent will load the adaptors found in the
> config/initial_adaptors file only of the checkpoints file doesn't exist. If
> the checkpoint file does exist, the initial_adaptors file is silently
> ignored. This is misleading and seems like a potential cause of confusion.
> Instead both files should both be checked for adaptors, first the checkpoint
> file, then the initial_adaptors file. This order would ensure that the same
> adaptor won't lose it's checkpoint info or get loaded twice if the agent is
> bounced.
> Also, the agents.log file does a good job of reporting which adaptors are
> registered, but it doesn't show where they were loaded from. This information
> would be useful to have if possible.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
