You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "Luke Chen (Jira)" <ji...@apache.org> on 2022/07/08 13:05:00 UTC
[jira] [Updated] (KAFKA-13996) log.cleaner.io.max.bytes.per.second cannot be changed dynamically
[ https://issues.apache.org/jira/browse/KAFKA-13996?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Luke Chen updated KAFKA-13996:
------------------------------
Fix Version/s: 3.3.0
> log.cleaner.io.max.bytes.per.second cannot be changed dynamically
> ------------------------------------------------------------------
>
> Key: KAFKA-13996
> URL: https://issues.apache.org/jira/browse/KAFKA-13996
> Project: Kafka
> Issue Type: Bug
> Components: config, core, log cleaner
> Affects Versions: 3.2.0
> Reporter: Tomonari Yamashita
> Assignee: Tomonari Yamashita
> Priority: Major
> Fix For: 3.3.0
>
>
> - log.cleaner.io.max.bytes.per.second cannot be changed dynamically using bin/kafka-configs.sh
> - Reproduction procedure:
> -# Create a topic with cleanup.policy=compact
> {code:java}
> bin/kafka-topics.sh --bootstrap-server localhost:9092 --create --replication-factor 1 --partitions 1 --topic my-topic --config cleanup.policy=compact --config cleanup.policy=compact --config segment.bytes=104857600 --config compression.type=producer
> {code}
> -# Change log.cleaner.io.max.bytes.per.second=10485760 using bin/kafka-configs.sh
> {code:java}
> bin/kafka-configs.sh --bootstrap-server localhost:9092 --entity-type brokers --entity-default --alter --add-config log.cleaner.io.max.bytes.per.second=10485760
> {code}
> -# Send enough messages(> segment.bytes=104857600) to activate Log Cleaner
> -# logs/log-cleaner.log, configuration by log.cleaner.io.max.bytes.per.second=10485760 is not reflected and Log Cleaner does not slow down (>= log.cleaner.io.max.bytes.per.second=10485760).
> {code:java}
> [2022-06-15 14:52:14,988] INFO [kafka-log-cleaner-thread-0]:
> Log cleaner thread 0 cleaned log my-topic-0 (dirty section = [39786, 81666])
> 3,999.0 MB of log processed in 2.7 seconds (1,494.4 MB/sec).
> Indexed 3,998.9 MB in 0.9 seconds (4,218.2 Mb/sec, 35.4% of total time)
> Buffer utilization: 0.0%
> Cleaned 3,999.0 MB in 1.7 seconds (2,314.2 Mb/sec, 64.6% of total time)
> Start size: 3,999.0 MB (41,881 messages)
> End size: 0.1 MB (1 messages)
> 100.0% size reduction (100.0% fewer messages)
> (kafka.log.LogCleaner)
> {code}
> - Problem cause:
> -- log.cleaner.io.max.bytes.per.second is used in Throttler in LogCleaner, however, it is only passed to Throttler at initialization time.
> --- https://github.com/apache/kafka/blob/4380eae7ceb840dd93fee8ec90cd89a72bad7a3f/core/src/main/scala/kafka/log/LogCleaner.scala#L107-L112
> -- Need to change Throttler configuration value at reconfigure() of LogCleaner.
> --- https://github.com/apache/kafka/blob/4380eae7ceb840dd93fee8ec90cd89a72bad7a3f/core/src/main/scala/kafka/log/LogCleaner.scala#L192-L196
> - A workaround is that restarting every broker adding log.cleaner.io.max.bytes.per.second to config/server.properties
--
This message was sent by Atlassian Jira
(v8.20.10#820010)