apache/iceberg · error · LockException

Interrupted while trying to find lock for table %s.%s

Error message

Interrupted while trying to find lock for table %s.%s

What it means

While polling showLocks to confirm the newly created lock, the waiting thread can be interrupted. The code restores the interrupt flag, logs a warning with the cause, and rethrows as a LockException wrapping the InterruptedException. This signals the commit's lock acquisition was cancelled by thread interruption rather than a metastore problem.

Source

Thrown at hive-metastore/src/main/java/org/apache/iceberg/hive/MetastoreLock.java:333

                    if (lockFound != null) {
                      lockInfo.lockId = lockFound.lockId;
                      lockInfo.lockState = lockFound.lockState;
                      LOG.info("Found lock {} by agentInfo {}", lockInfo, agentInfo);
                      return;
                    }
                  }

                  throw new LockException(
                      "Failed to find lock for table %s.%s", databaseName, tableName);
                } catch (InterruptedException e) {
                  Thread.currentThread().interrupt();
                  interrupted.set(true);
                  LOG.warn(
                      "Interrupted while trying to find lock for table {}.{}",
                      databaseName,
                      tableName,
                      e);
                  throw new LockException(
                      e,
                      "Interrupted while trying to find lock for table %s.%s",
                      databaseName,
                      tableName);
                }
              } catch (InterruptedException e) {
                Thread.currentThread().interrupt();
                interrupted.set(true);
                LOG.warn(
                    "Interrupted while creating lock on table {}.{}", databaseName, tableName, e);
                throw new LockException(
                    e, "Interrupted while creating lock on table %s.%s", databaseName, tableName);
              }
            },
            LockException.class);

    // This should be initialized always, or exception should be thrown.
    LOG.debug("Lock {} created for table {}.{}", lockInfo, databaseName, tableName);

View on GitHub (pinned to 86d9c8fc54)

Solutions

  1. Identify and remove the source of interruption (task cancellation, executor shutdown) before committing.
  2. Retry the commit from a non-interrupted thread; the interrupt flag is preserved so the caller must decide to retry.
  3. If committing in a thread pool, ensure shutdown/cancellation logic drains in-flight commits or shields commit threads from interruption.
  4. Check logs for the logged InterruptedException cause to find who interrupted the thread.

Example fix

// before
executor.shutdownNow(); // interrupts in-flight commit threads
// after
executor.shutdown();
executor.awaitTermination(5, TimeUnit.MINUTES); // let commits finish before forcing shutdown
Defensive patterns

Strategy: try-catch

Validate before calling

if (Thread.currentThread().isInterrupted()) {
  throw new CancellationException("Refusing to commit on interrupted thread");
}

Try / catch

try {
  table.commit(apply);
} catch (LockException e) {
  if (e.getCause() instanceof InterruptedException) {
    Thread.currentThread().interrupt(); // preserve flag; treat as cancellation
    return;
  }
  throw e;
}

Prevention

When it happens

Trigger: Calling commit on a Hive-catalog table from a thread that gets interrupted during the lock-acquisition polling loop — e.g. executor shutdown, task cancellation, or query kill (Spark/Flink cancel) interrupting worker threads.

Common situations: Spark stage cancellation or speculative-execution kills interrupting the commit thread; application shutdown (SIGTERM to the executor) interrupting scheduler threads; manual Future.cancel(true) on the commit task.

Related errors

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/224f87aaef881cf5. Report an issue: GitHub.