[
https://issues.apache.org/jira/browse/LOG4J2-1044?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14588655#comment-14588655
]
Siva Kotipalli commented on LOG4J2-1044:
----------------------------------------
Thanks for the fix, i had verified the source, but it has few issues:
1. When logger is shutdown before it reaches the batchSize, the events will be
lost.
2. As i understand from documentation, the delay seconds is used when batch
size is not reached, but as per current implementation the delay second check
is included under send method, which may have issues, for example batch size is
1000 and delay secs is 10 mins, assuming we got 990 events continuously and it
may take another 1 hr for receiving the next event, then the transfer of the
data will not be done for 1 hr even though the delay is 10 mins.
I would prefer to fix these issues, can i provide patch for the same.
> batchSize doesnt work for flume avro manager
> --------------------------------------------
>
> Key: LOG4J2-1044
> URL: https://issues.apache.org/jira/browse/LOG4J2-1044
> Project: Log4j 2
> Issue Type: Question
> Components: Flume Appender
> Affects Versions: 2.2
> Reporter: Siva Kotipalli
> Assignee: Ralph Goers
> Fix For: 2.4
>
>
> The flume appender documentation mention about batch size which states that
> "Specifies the number of events that should be sent as a batch. The default
> is 1.". The batchSize doesnt work and always defaults to one.
> Its noticed that method FlumeAvroManager.send(final BatchEvent events) is
> never referred by appender.
> Kindly let us know what is the correct config for batch size, the provided
> config is
> <Flume name="flume_avro" compress="false" batchSize="5">...
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]