[IOTDB-3732] [same folder but different rpc port]- "io task meets error" is not handled correctly

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 0.13.0, 0.12.5
• Fix Version/s: 1.0.0
• Labels: pull-request-available
• Assignee: Haonan Hou
• Reporter: FengQingxin
• Created: 04/Jul/22 02:25
• Updated: 28/Nov/22 06:35
• Resolved: 14/Nov/22 05:31

Description

Hi 

If I restart a iotdb istance with default config item first, then I changed rpc port to 6668 and start another iotdb instance in this folder.

Does the following question still exist?

 

 

 

 

↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓cloned from  issue 3306↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓

When restarting IoTDB and flushing the wal, the io task meets error. 

It throw a RuntimeException, however, this exception is not catched by TsFileProcessor, TsFileProcessor still ends the file, which causes a corrupt tsfile.

 

2022-04-21 22:29:19,461 [pool-2-IoTDB-Flush-SubTask-2] ERROR o.a.i.d.e.f.MemTableFlushTask:297 - Storage group 2 memtable PrimitiveMemTable

{planIndex=[0,0]}

, io task meets error. 
java.io.IOException: Bytes written is inconsistent with the size of data: 57 != 45
    at org.apache.iotdb.tsfile.write.chunk.ChunkWriterImpl.writeAllPagesOfChunkToTsFile(ChunkWriterImpl.java:459)
    at org.apache.iotdb.tsfile.write.chunk.ChunkWriterImpl.writeToFileWriter(ChunkWriterImpl.java:323)
    at org.apache.iotdb.db.engine.flush.MemTableFlushTask.lambda$new$0(MemTableFlushTask.java:290)
    at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
    at java.util.concurrent.FutureTask.run(FutureTask.java:266)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at java.lang.Thread.run(Thread.java:748)

 

2022-04-21 22:29:24,797 [Thread-4] INFO  o.a.i.d.e.f.MemTableFlushTask:150 - Storage group 2 memtable PrimitiveMemTable{planIndex=[0,0]}

flushing a memtable has finished! Time consumption: 5448ms 

Issue Links

• is a clone of: IOTDB-3306 "io task meets error" is not handled correctly
• is fixed by: IOTDB-4402 Avoid multiple IoTDB processes accessing same data dir
• relates to: IOTDB-4772 error would occurs when starting datanode 2 times and all nodes seems normal with show cluster