Yiqun Lin created HDFS-12565:
--------------------------------
Summary: Ozone: Put key operation concurrent executes failed on
Windows
Key: HDFS-12565
URL: https://issues.apache.org/jira/browse/HDFS-12565
Project: Hadoop HDFS
Issue Type: Sub-task
Components: ozone
Affects Versions: HDFS-7240
Reporter: Yiqun Lin
When creating a batch size of key under specified bucket, then the error
happens on Windows. This error was found by executing
{{TestOzoneShell#testListKey()}}.
{noformat}
org.apache.hadoop.scm.container.common.helpers.StorageContainerException:
org.apache.hadoop.scm.container.common.helpers.StorageContainerException:
Invalid write size found. Size: 1768160 Expected: 10
at
org.apache.hadoop.scm.storage.ContainerProtocolCalls.validateContainerResponse(ContainerProtocolCalls.java:373)
at
org.apache.hadoop.scm.storage.ContainerProtocolCalls.writeChunk(ContainerProtocolCalls.java:175)
at
org.apache.hadoop.scm.storage.ChunkOutputStream.writeChunkToContainer(ChunkOutputStream.java:224)
at
org.apache.hadoop.scm.storage.ChunkOutputStream.close(ChunkOutputStream.java:154)
at
org.apache.hadoop.ozone.client.io.ChunkGroupOutputStream$ChunkOutputStreamEntry.close(ChunkGroupOutputStream.java:265)
at
org.apache.hadoop.ozone.client.io.ChunkGroupOutputStream.close(ChunkGroupOutputStream.java:174)
at
org.apache.hadoop.ozone.client.io.OzoneOutputStream.close(OzoneOutputStream.java:58)
at
org.apache.hadoop.ozone.web.storage.DistributedStorageHandler.commitKey(DistributedStorageHandler.java:405)
at
org.apache.hadoop.ozone.web.handlers.KeyHandler$2.doProcess(KeyHandler.java:196)
at
org.apache.hadoop.ozone.web.handlers.KeyProcessTemplate.handleCall(KeyProcessTemplate.java:91)
at
org.apache.hadoop.ozone.web.handlers.KeyHandler.putKey(KeyHandler.java:199)
{noformat}
The related codes(ChunkUtils#writeData):
{code}
public static void writeData(File chunkFile, ChunkInfo chunkInfo,
byte[] data) throws
StorageContainerException, ExecutionException, InterruptedException,
NoSuchAlgorithmException {
...
try {
file =
AsynchronousFileChannel.open(chunkFile.toPath(),
StandardOpenOption.CREATE,
StandardOpenOption.WRITE,
StandardOpenOption.SPARSE,
StandardOpenOption.SYNC);
lock = file.lock().get();
if (chunkInfo.getChecksum() != null &&
!chunkInfo.getChecksum().isEmpty()) {
verifyChecksum(chunkInfo, data, log);
}
int size = file.write(ByteBuffer.wrap(data), chunkInfo.getOffset()).get();
if (size != data.length) { <===== error was thrown
log.error("Invalid write size found. Size:{} Expected: {} ", size,
data.length);
throw new StorageContainerException("Invalid write size found. " +
"Size: " + size + " Expected: " + data.length, INVALID_WRITE_SIZE);
}
...
{code}
But if we only put one single key, it runs well.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-dev-h...@hadoop.apache.org
