[
https://issues.apache.org/jira/browse/FLINK-24733?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17437403#comment-17437403
]
Yufan Sheng commented on FLINK-24733:
-------------------------------------
[~shengzhi] Tks for your fix. I'm digging into the Pulsar client API for making
sure this fix is OK.
I think you may need to add another fix for the ordered reader.
> Data loss in pulsar source when using shared mode
> -------------------------------------------------
>
> Key: FLINK-24733
> URL: https://issues.apache.org/jira/browse/FLINK-24733
> Project: Flink
> Issue Type: Bug
> Components: Connectors / Pulsar
> Affects Versions: 1.14.0
> Reporter: wang shengzhi
> Priority: Major
> Labels: pull-request-available
>
> Description：
> Noticed that when using Flink-connector-pulsar with Shared mode (default
> mode), some messages are lost.
> Details：
> When a topic partition does not receive message for more than 10s
> (pulsar.source.maxFetchTime), the next coming message will be dropped by
> source.
> In stream applications, this situation which there is no data for more than
> 10s is very common. So this bug will cause data loss.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
