[
https://issues.apache.org/jira/browse/CHUKWA-523?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12921903#action_12921903
]
Eric Yang commented on CHUKWA-523:
----------------------------------
It looks like the reload command has been send through
AgentControlSocketListener -> HttpConnector, but ChukwaHttpSender is stuck on
the original list. The maximum retries is very large in reliablySend by
default (144000 Retries). When ChukwaHttpSender reaches maximum retry count
with the original list, it will throw IOException("no collectors") to
HttpConnector. HttpConnector will cause agent to shutdown due to no collector
to try. The updated collector list is never really deployed to
ChukwaHttpSender.
Collector list change should be notifying ChukwaHttpSender as soon as
reloadConfiguration is called.
> Reload collectors doesn't work
> ------------------------------
>
> Key: CHUKWA-523
> URL: https://issues.apache.org/jira/browse/CHUKWA-523
> Project: Chukwa
> Issue Type: Bug
> Components: data collection
> Affects Versions: 0.4.0
> Environment: Java 6, Mac OS X 10.6
> Reporter: Eric Yang
>
> 1) Start the agent with some "X" collectors.
> 2) Update the conf/collectors file with "Y" collector entry and removing
> "X" collector entry.
> 3) Doing telnet to port 9093
> 4) Firing command "reloadcollectors".
>
> After reloading the collectors, agent responses as "OK reloadCollectors
> done". But agents logs shows trying to connect to old collector "X" and does
> not connect to new collector.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
