[
https://issues.apache.org/jira/browse/KAFKA-1566?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14084788#comment-14084788
]
Cosmin Lehene commented on KAFKA-1566:
--------------------------------------
Somehow related: KAFKA_LOG_DIR would be better than setting KAFKA_LOG4J_OPTS.
kafka-run-class.sh would check for this
{code}
LOG_DIR=${KAFKA_LOG_DIR:-"$base_dir/logs"}
{code}
Perhaps a new issue may be useful (otherwise I could just add them to this
patch).
> Kafka environment configuration (kafka-env.sh)
> ----------------------------------------------
>
> Key: KAFKA-1566
> URL: https://issues.apache.org/jira/browse/KAFKA-1566
> Project: Kafka
> Issue Type: Improvement
> Components: tools
> Reporter: Cosmin Lehene
> Assignee: Cosmin Lehene
> Labels: newbie
> Fix For: 0.8.2, 0.9.0
>
>
> It would be useful (especially for automated deployments) to have an
> environment configuration file that could be sourced from the launcher files
> (e.g. kafka-run-server.sh).
> This is how this could look like kafka-env.sh
> {code}
> export KAFKA_JVM_PERFORMANCE_OPTS="-XX:+UseCompressedOops
> -XX:+DisableExplicitGC -Djava.awt.headless=true \ -XX:+UseG1GC
> -XX:PermSize=48m -XX:MaxPermSize=48m -XX:MaxGCPauseMillis=20
> -XX:InitiatingHeapOccupancyPercent=35' %>"
> export KAFKA_HEAP_OPTS="'-Xmx1G -Xms1G' %>"
> export KAFKA_LOG4J_OPTS="-Dkafka.logs.dir=/var/log/kafka"
> {code}
> kafka-server-start.sh
> {code}
> ...
> source $base_dir/config/kafka-env.sh
> ...
> {code}
> This approach is consistent with Hadoop and HBase. However the idea here is
> to be able to set these values in a single place without having to edit
> startup scripts.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
