[
https://issues.apache.org/jira/browse/FLINK-35890?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Leonard Xu reassigned FLINK-35890:
----------------------------------
Assignee: sanqingleo
> Select only the increment but still enter the SNAPSHOT,lead
> ------------------------------------------------------------
>
> Key: FLINK-35890
> URL: https://issues.apache.org/jira/browse/FLINK-35890
> Project: Flink
> Issue Type: Bug
> Components: Flink CDC
> Affects Versions: cdc-3.1.1
> Reporter: sanqingleo
> Assignee: sanqingleo
> Priority: Major
> Labels: pull-request-available
> Attachments: image-2024-07-24-16-47-19-258.png
>
>
> Hi, when I was using Flink CDC Postgre, I chose 'debezium.snapshot.mode' =
> 'never' to skip the snapshot phase, but the code still entered the snapshot
> branch, and this log was printed, 'Database snapshot phase can 't perform
> checkpoint, acquired Checkpoint lock.'
> (Since our business creates the task first and then writes the data), the
> task cannot perform ckp, and the status is always stuck in in progress. After
> the timeout, the task restarts.
> Normally, if you only choose incremental data synchronization, you should not
> enter this branch.
>
> The failure of ckp does not mean that the first one fails. After multiple ckp
> are successfully performed, it will then get stuck, causing a timeout and
> restarting the task. This is a problem that must occur. After it appeared in
> production, it also recurred many times in local testing.
>
> Then we test, if there is data flowing in, the task status will be normal.
>
> sql and conf like this
> !image-2024-07-24-16-47-19-258.png!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
