[ https://issues.apache.org/jira/browse/KAFKA-6701?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Dong Lin resolved KAFKA-6701. ----------------------------- Resolution: Fixed The issue appears to have been fixed in https://issues.apache.org/jira/browse/KAFKA-5163. More specifically, https://issues.apache.org/jira/browse/KAFKA-5163 added method `Log.renameDir()` and this method will grab the per-log lock before making modification to the log's directory etc. > synchronize Log modification between delete cleanup and async delete > -------------------------------------------------------------------- > > Key: KAFKA-6701 > URL: https://issues.apache.org/jira/browse/KAFKA-6701 > Project: Kafka > Issue Type: Bug > Reporter: Sumant Tambe > Assignee: Sumant Tambe > Priority: Major > > Kafka broker crashes without any evident disk failures > From [~becket_qin]: This looks a bug in kafka when topic deletion and log > retention cleanup happen at the same time, the log retention cleanup may see > ClosedChannelException after the log has been renamed for async deletion. > The root cause is that the topic deletion should have set the isClosed flag > of the partition log to true and the retention should not bother to do the > old log segments deletion when the log is closed. -- This message was sent by Atlassian JIRA (v7.6.3#76005)