log IT

Morning,

If you have a Apache Kafka version which is below 1.0.0 and you don’t have garbage collect rotate as shown here:

with:

-Xloggc:/opt/kafka/bin/../logs/kafkaServer-gc.log -verbose:gc -XX:+PrintGCDetails -XX:+PrintGCDateStamps -XX:+PrintGCTimeStamps -XX:+UseGCLogFileRotation -XX:NumberOfGCLogFiles=10 -XX:GCLogFileSize=100M

without:

-Xloggc:/opt/kafka/bin/../logs/kafkaServer-gc.log -verbose:gc -XX:+PrintGCDetails -XX:+PrintGCDateStamps -XX:+PrintGCTimeStamps

One option is to modify the parameters in order to include them in the process that starts but this involves also restarting the services.

However, you can use also logrotate daemon with the following configuration, and i will put the block that you need to add on hiera using:

logrotate::rule:
 'kafka_gc':
   path: '/opt/kafka/logs/kafkaServer-gc.log'
   copytruncate: true
   rotate_every: 'day'
   compress: true
   missingok: true
   su: true
   su_owner: 'kafka'
   su_group: 'kafka'
   ifempty: false
   size: '50M'
   maxsize: '50M'
   rotate: 5

Or if you want to write it in a class, it should look like

$version = lookup('kafka::version')
        if ($_role =~ /\Akafka/) and ($version != '1.0.0') {
            logrotate::rule { 'kafkagc_logs':
             path => '/opt/kafka/logs/kafkaServer-gc.log',
             copytruncate => true,
             rotate => 5,
             rotate_every => 'daily',
             missingok => true,
             ifempty => false,
             su => true,
             su_owner => 'kafka',
             su_group => 'kafka',
             size => '50M',
             maxsize => '50M',
        }                                              

Cheers!