[
https://issues.apache.org/jira/browse/ACCUMULO-4409?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16203005#comment-16203005
]
Josh Elser commented on ACCUMULO-4409:
--------------------------------------
bq. If you can reproduce with a full stack trace, that might help.
That was all that was given to me by Maven. On a re-run of the {{mvn package}},
it passed. If you can't reproduce, I guess we just close again. Very strange.
> Create a custom log4j AccumuloMonitorAppender
> ---------------------------------------------
>
> Key: ACCUMULO-4409
> URL: https://issues.apache.org/jira/browse/ACCUMULO-4409
> Project: Accumulo
> Issue Type: Sub-task
> Reporter: Christopher Tubbs
> Assignee: Christopher Tubbs
> Labels: release_notes
> Fix For: 2.0.0
>
> Time Spent: 6h
> Remaining Estimate: 0h
>
> The current AsyncSocketAppender was an improvement over the older log4j
> configuration options, but it did not fully satisfy the description in
> ACCUMULO-2343.
> What we really need is an appender which can be managed by configuration
> alone, without using code to reload it with new system properties when the
> location of the monitor changes.
> So, rather than having a ZK watcher reload the log4j system, the appender
> itself should internally detect the relocation of the monitor and adjust its
> behavior accordingly. That way, the configuration files do not need to refer
> to system properties for the hostname/port of the monitor, and there is no
> need to reload the log4j logging when things change.
> This will allow us to remove more custom log4j code, which gets us closer to
> users having the ability to use another implementation for the slf4j API
> (though this feature of logging to the monitor might not be available for
> other slf4j implementations, unless they support using log4j appenders).
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
