You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "Daniel Urban (Jira)" <ji...@apache.org> on 2025/03/06 12:28:00 UTC
[jira] [Updated] (KAFKA-18929) Log a warning when a time retention based segment has a future timestamp
[ https://issues.apache.org/jira/browse/KAFKA-18929?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Daniel Urban updated KAFKA-18929:
---------------------------------
Description:
Time based retention can be blocked when a producer sends a future timestamp in a record. This can cause log dirs to run out of free space.
This seems to be a frequent issue which can be mitigated by using the configurations added in [KIP-937|[https://cwiki.apache.org/confluence/display/KAFKA/KIP-937%3A+Improve+Message+Timestamp+Validation]] but that is a somewhat drastic measure and also requires users to explicitly configure the timestamp limits.
Adding a warning log when the log cleaning encounters a future timestamp can improve catching this issue earlier.
was:
Time based retention can be blocked when a producer sends a future timestamp in a record. This can cause log dirs to run out of free space.
This seems to be a frequent issue which can be mitigated by using the configurations added in [KIP-937|[https://cwiki.apache.org/confluence/display/KAFKA/KIP-937%3A+Improve+Message+Timestamp+Validation],] but that is a somewhat drastic measure and also requires users to explicitly configure the timestamp limits.
Adding a warning log when the log cleaning encounters a future timestamp can improve catching this issue earlier.
> Log a warning when a time retention based segment has a future timestamp
> ------------------------------------------------------------------------
>
> Key: KAFKA-18929
> URL: https://issues.apache.org/jira/browse/KAFKA-18929
> Project: Kafka
> Issue Type: Improvement
> Components: log, log cleaner, logging
> Reporter: Daniel Urban
> Assignee: Daniel Urban
> Priority: Major
>
> Time based retention can be blocked when a producer sends a future timestamp in a record. This can cause log dirs to run out of free space.
> This seems to be a frequent issue which can be mitigated by using the configurations added in [KIP-937|[https://cwiki.apache.org/confluence/display/KAFKA/KIP-937%3A+Improve+Message+Timestamp+Validation]] but that is a somewhat drastic measure and also requires users to explicitly configure the timestamp limits.
>
> Adding a warning log when the log cleaning encounters a future timestamp can improve catching this issue earlier.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)