Lu Niu created FLINK-36582:
------------------------------

             Summary: sql client hang when querying a parquet table
                 Key: FLINK-36582
                 URL: https://issues.apache.org/jira/browse/FLINK-36582
             Project: Flink
          Issue Type: Improvement
          Components: Formats (JSON, Avro, Parquet, ORC, SequenceFile), Table 
SQL / Runtime
    Affects Versions: 1.18.1
            Reporter: Lu Niu
         Attachments: sql_client_jstack_1

flinksql client will hang when execute a query:
{code:java}
select * from table_foo where dt = "2024-10-21" {code}
table_foo is a partitioned table.  "2024-10-21"  partition points to a dir on 
s3 containing 6000 parquet files.

jstack shows the parquet-get-table-statistic-worker-threads are the running 
threads. 

the issue can be mitigated after:
{code:java}
table.optimizer.source.report-statistics-enabled: false 
source.report-statistics: NONE   {code}
However not sure about downside disabling this



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to