[jira] [Commented] (IOTDB-573) Construct AbstractIoTDBThread and catch all Exception

["Jialin Qiao (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IOTDB-573?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17072719#comment-17072719 ] 

Jialin Qiao commented on IOTDB-573:
-----------------------------------

For example, In StorageEngine

 

List<StorageGroupMNode> sgNodes = MManager.getInstance().getAllStorageGroupNodes();
List<Future> futures = new ArrayList<>();
for (StorageGroupMNode storageGroup : sgNodes) {
 futures.add(recoveryThreadPool.submit((Callable<Void>) () -> {
 StorageGroupProcessor processor = new StorageGroupProcessor(systemDir,
 storageGroup.getFullPath(), fileFlushPolicy);
 processor.setDataTTL(storageGroup.getDataTTL());
 processorMap.put(storageGroup.getFullPath(), processor);
 logger.info("Storage Group Processor {} is recovered successfully",
 storageGroup.getFullPath());
 return null;
 }));
}
for (Future future : futures) {
 try {
 future.get();
 } catch (InterruptedException | ExecutionException e) {
 throw new StorageEngineFailureException("StorageEngine failed to recover.", e);
 }
}

 

We will get a NullPointerException without knowing the stack information if the constructor StorageGroupProcessor throws a NullPointerException.

 

> Construct  AbstractIoTDBThread and catch all Exception
> ------------------------------------------------------
>
>                 Key: IOTDB-573
>                 URL: https://issues.apache.org/jira/browse/IOTDB-573
>             Project: Apache IoTDB
>          Issue Type: Improvement
>          Components: Core/Engine
>            Reporter: Jialin Qiao
>            Priority: Major
>
> Hi, in current IoTDB codes, threads are created without any constraints. RuntimeExceptions are not caught in the run method. Therefore, we usually miss the stack information and only get a NullPointerException in the parent thread.
>  
> Therefore, we need a base thread AbstractIoTDBThread that manages all threads. In the IoTDBThead, all exceptions are caught and logged. Here is an example.
>  
> import java.util.concurrent.Callable;
> import org.slf4j.Logger;
> import org.slf4j.LoggerFactory;
> public abstract class AbstractIoTDBThread<T> implements Callable<T> {
>  protected Logger logger = null;
>  public AbstractIoTDBThread() {
>  initLogger();
>  }
>  @Override
>  public T call() throws Exception {
>  try {
>  return internalRun();
>  } catch (Exception e) {
>  logger.error("Meet error in sub thread", e);
>  throw e;
>  }
>  }
>  abstract void initLogger();
>  abstract T internalRun();
> }
> class MyThread extends AbstractIoTDBThread<Long> {
>  @Override
>  void initLogger() {
>  logger = LoggerFactory.getLogger(MyThread.class);
>  }
>  @Override
>  Long internalRun() {
>  return 0L;
>  }
> }
>  
> This issue contains two tasks:
> (1) Create and enrich AbstractIoTDBThread
> (2) Make all threads in IoTDB to extend AbstractIoTDBThread.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)