MingHui Luo created HDFS-16970:
----------------------------------
Summary: EC: client copy wrong buffer from decode ouput during
pread
Key: HDFS-16970
URL: https://issues.apache.org/jira/browse/HDFS-16970
Project: Hadoop HDFS
Issue Type: Bug
Components: dfsclient, ec, erasure-coding
Affects Versions: 3.3.4
Reporter: MingHui Luo
Fix For: 3.4.0, 3.2.5, 3.3.5
When dfsStripedInputStream do pread from a striped block group and read
internal block timeout, so will read parity block for decode and fill original
chunk buffer with decoded data.
Here try to fill original chunk buffer with decoded data, but get wrong data.
The reason is that
1.original chunk buffer already read some bytes before timeout from blockReader
2.chunkBytebuffer's slice always fill begin 0 position of decodeByteBuffer
slice bytebuffer will fill from wrong decodeByteBuffer position, so will get
wrong data from pread.
{code:java}
23/03/21 06:31:11 WARN [StripedRead-24] DFSClient: Exception while reading from
BP-xxx:blk_-9xxx_xxx of file_xxx from
DatanodeInfoWithStorage[10.xxx.xx.xx:50010,DS-xxx,DISK]
java.net.SocketTimeoutException: 10000 millis timeout while waiting for channel
to be ready for read. ch : java.nio.channels.SocketChannel[connected
local=/10.xxx.xx.xx:51426 remote=/10.xxx.xx.xx:50010]
at
org.apache.hadoop.net.SocketIOWithTimeout.doIO(SocketIOWithTimeout.java:164)
at org.apache.hadoop.net.SocketInputStream.read(SocketInputStream.java:161)
at
org.apache.hadoop.hdfs.protocol.datatransfer.PacketReceiver.readChannelFully(PacketReceiver.java:256)
at
org.apache.hadoop.hdfs.protocol.datatransfer.PacketReceiver.doReadFully(PacketReceiver.java:207)
at
org.apache.hadoop.hdfs.protocol.datatransfer.PacketReceiver.doRead(PacketReceiver.java:134)
at
org.apache.hadoop.hdfs.protocol.datatransfer.PacketReceiver.receiveNextPacket(PacketReceiver.java:102)
at
org.apache.hadoop.hdfs.client.impl.BlockReaderRemote.readNextPacket(BlockReaderRemote.java:221)
at
org.apache.hadoop.hdfs.client.impl.BlockReaderRemote.read(BlockReaderRemote.java:201)
at
org.apache.hadoop.hdfs.ByteBufferStrategy.readFromBlock(ReaderStrategy.java:180)
at
org.apache.hadoop.hdfs.ByteBufferStrategy.readFromBlock(ReaderStrategy.java:172)
at org.apache.hadoop.hdfs.StripeReader.readToBuffer(StripeReader.java:240)
at
org.apache.hadoop.hdfs.StripeReader.lambda$readCells$0(StripeReader.java:286)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.util.concurrent.Executors$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) {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-dev-h...@hadoop.apache.org
