[
https://issues.apache.org/jira/browse/FLINK-37508?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17936711#comment-17936711
]
Hongshun Wang commented on FLINK-37508:
---------------------------------------
[~Leonard] , I'd like to do it.
> Postgres CDC Jdbc query should use debezium.snapshot.fetch.size rather than
> debezium.query.fetch.size.
> -------------------------------------------------------------------------------------------------------
>
> Key: FLINK-37508
> URL: https://issues.apache.org/jira/browse/FLINK-37508
> Project: Flink
> Issue Type: Improvement
> Components: Flink CDC
> Affects Versions: cdc-3.3.0
> Reporter: Hongshun Wang
> Priority: Major
> Fix For: cdc-3.4.0
>
>
> In debezium postgres connector, [{\{snapshot.fetch.size
> }}|https://debezium.io/documentation//reference/2.7/connectors/postgresql.html#postgresql-property-snapshot-fetch-size]specifies
> the maximum number of rows in a batch(defalut value is 10240).
> {{However, currently in pg cdc, use query.size(which is not a param of
> debezium postgres connector, the defalut value is 0, meaning read a without a
> fetch size). If the chunk size is huge, will OOM(directly)}}
> {code:java}
> PostgresQueryUtils.readTableSplitDataStatement(
> jdbcConnection,
> selectSql,
> snapshotSplit.getSplitStart() == null,
> snapshotSplit.getSplitEnd() == null,
> snapshotSplit.getSplitStart(),
> snapshotSplit.getSplitEnd(),
> snapshotSplit.getSplitKeyType().getFieldCount(),
> connectorConfig.getQueryFetchSize()); {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
