HIVE-29268: Iceberg: Close catalog in Catalogs.loadTable() to fix resource leak from open ResolvingFileIO

[difin]

What changes were proposed in this pull request?

Added try-with-resources for a catalog that is loaded in Catalogs.loadTable().

Why are the changes needed?

Described in the description of HIVE-29268.

Does this PR introduce any user-facing change?

No

How was this patch tested?

Manual check on the local Hive Docker instance.

[okumin]

This file is a copy of apache/iceberg/mr/.../Catalogs.java, and the upstream version does not have the equivalent of this change. Why does Apache Iceberg justify the current implementation? Or it is still a problem on upstream, but no MapReduce users are interested in the warning message.

The current implementation is likely to work. For future reference, I'm also curious about what the to-be solution is.

[deniskuzZ]

can we avoid logical fallback and close resource when required?

private static Table loadTable(Configuration conf, String tableIdentifier, String tableLocation,
                               String catalogName) {
    Optional<Catalog> catalog = loadCatalog(conf, catalogName);

    if (catalog.isPresent()) {
        Preconditions.checkArgument(tableIdentifier != null, "Table identifier not set");
        try {
            return catalog.get().loadTable(TableIdentifier.parse(tableIdentifier));
        } finally {
            if (catalog.get() instanceof Closeable closeable) {
              IOUtils.closeQuietly(closeable);
            }
        }
    }
    ....
}

[okumin]

I guess Java 21 provides a smarter way to express this cast.
https://sonarcloud.io/project/issues?id=apache_hive&pullRequest=6131&issueStatuses=OPEN,CONFIRMED&sinceLeakPeriod=true

[difin]

I found that this proposed change resolves the warning for the unclosed ResolvingFileIO instance in the logs, but it break the functionality, because it closes HiveRESTCatalogClient connection to the RESTCatalog server and causes below exceptions when Iceberg tries to commit to an iceberg table.

2025-10-15 13:29:33 2025-10-15T17:29:33,844 ERROR [HiveServer2-Background-Pool: Thread-159] hive.HiveIcebergStorageHandler: Error while trying to commit job: job_17605493726650_0001, starting rollback changes for table: ice_rest.ice_orc2
2025-10-15 13:29:33 java.lang.IllegalStateException: Connection pool shut down
2025-10-15 13:29:33     at org.apache.hc.core5.util.Asserts.check(Asserts.java:38) ~[httpcore5-5.3.1.jar:5.3.1]
2025-10-15 13:29:33     at org.apache.hc.core5.pool.StrictConnPool.lease(StrictConnPool.java:175) ~[httpcore5-5.3.1.jar:5.3.1]
...

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud