apache/iceberg · error · NoSuchTableException

Cannot find Glue table %s after refresh, maybe another proce

Error message

Cannot find Glue table %s after refresh, maybe another process deleted it or revoked your access permission

What it means

GlueTableOperations.doRefresh throws NoSuchTableException when a refresh finds the Glue table missing (getGlueTable returns null) even though the operation previously knew a metadata location. This indicates the table was deleted underneath the running operation or access was revoked, so the current state cannot be refreshed.

Source

Thrown at aws/src/main/java/org/apache/iceberg/aws/glue/GlueTableOperations.java:134

    }
    return fileIO;
  }

  @Override
  protected String tableName() {
    return fullTableName;
  }

  @Override
  protected void doRefresh() {
    String metadataLocation = null;
    Table table = getGlueTable();
    if (table != null) {
      checkIfTableIsIceberg(table, tableName());
      metadataLocation = table.parameters().get(METADATA_LOCATION_PROP);
    } else {
      if (currentMetadataLocation() != null) {
        throw new NoSuchTableException(
            "Cannot find Glue table %s after refresh, "
                + "maybe another process deleted it or revoked your access permission",
            tableName());
      }
    }

    refreshFromMetadataLocation(metadataLocation);
  }

  @Override
  protected void doCommit(TableMetadata base, TableMetadata metadata) {
    CommitStatus commitStatus = CommitStatus.FAILURE;
    RetryDetector retryDetector = new RetryDetector();

    String newMetadataLocation = null;
    boolean glueTempTableCreated = false;
    try {
      glueTempTableCreated = createGlueTempTableIfNecessary(base, metadata.location());

View on GitHub (pinned to 86d9c8fc54)

Solutions

1. Check whether another job or user dropped the Glue table (CloudTrail for DeleteTable events) and stop the drop or recreate the table.
2. Verify Lake Formation / IAM grants still allow the current role to read the table.
3. Recreate the table from its metadata if the deletion was accidental, then re-open the Iceberg table reference.
4. Coordinate table lifecycle across pipelines — never drop tables while readers hold live references.

Example fix

// before
Table table = catalog.loadTable(ident);
table.refresh(); // throws if table was deleted concurrently

// after
Table table = catalog.loadTable(ident);
if (!catalog.tableExists(ident)) {
  LOG.error("Table {} was deleted externally, aborting job", ident);
  return;
}
table.refresh();

TableIdentifier ident = TableIdentifier.of("sales", "events");
if (!catalog.tableExists(ident)) {
  throw new IllegalStateException("Table " + ident + " no longer exists in Glue");
}

try {
  table.refresh();
} catch (NoSuchTableException e) {
  LOG.error("Glue table {} was deleted or access revoked mid-job; aborting", tableName);
  throw e;
}

When it happens

Trigger: Calling table.refresh() or any operation triggering refresh (reads, commits) after another process/job issued Glue DeleteTable, or after IAM permissions on the table were revoked mid-session.

Common situations: Concurrent Spark/Flink jobs where one drops the table while another holds an open reference; retention jobs cleaning up 'old' tables still in use; Lake Formation permission changes removing access; long-running streaming queries whose underlying table was deleted.

Understand the failure class

Background: "Not found" and "does not exist" errors: why "Task not found", "No such folder", and "Can't find" fire when a lookup comes back empty — this error's family across 14 libraries.

Related errors

• Cannot commit %s due to unexpected exception
• Fail to acquire lock %s to commit new metadata at %s
• Cannot commit %s because base metadata location '%s' is not
• Cannot commit %s because Glue detected concurrent update
• Optimistic locking is not available in the environment. Usin

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/889603460e6d6d65. Report an issue: GitHub.