stanley shi created HDFS-6522:
---------------------------------
Summary: DN will try to append to non-existent replica if the
datanode has out-dated block
Key: HDFS-6522
URL: https://issues.apache.org/jira/browse/HDFS-6522
Project: Hadoop HDFS
Issue Type: Bug
Affects Versions: 2.2.0
Reporter: stanley shi
My environment: HA cluster with 4 datanodes;
Here're the steps to reproduce:
1. put one file (one block) to hdfs with repl=3; assume dn1, dn2, dn3 has block
for this file; dn4 don't have the block;
2. stop the dn1;
4. append content to the file 100 times;
5. close dn1 and start dn4;
6. append content to the file 100 times again;
Check the datanode log on dn1, many of this log will show {quote}
2014-06-12 12:07:04,442 INFO org.apache.hadoop.hdfs.server.datanode.DataNode:
opWriteBlock BP-1649188734-10.37.7.142-1398844098971:blk_1073742928_61304
received exception org.apache.hadoop.hdfs.s
erver.datanode.ReplicaNotFoundException: Cannot append to a non-existent
replica BP-1649188734-10.37.7.142-1398844098971:blk_1073742928_61304
2014-06-12 12:07:04,442 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode:
hdsh145.lss.emc.com:50010:DataXceiver error processing WRITE_BLOCK operation
src: /10.37.7.146:55594 dest: /10.37.7
.145:50010
org.apache.hadoop.hdfs.server.datanode.ReplicaNotFoundException: Cannot append
to a non-existent replica
BP-1649188734-10.37.7.142-1398844098971:blk_1073742928_61304
at
org.apache.hadoop.hdfs.server.datanode.fsdataset.impl.FsDatasetImpl.getReplicaInfo(FsDatasetImpl.java:392)
at
org.apache.hadoop.hdfs.server.datanode.fsdataset.impl.FsDatasetImpl.append(FsDatasetImpl.java:527)
at
org.apache.hadoop.hdfs.server.datanode.fsdataset.impl.FsDatasetImpl.append(FsDatasetImpl.java:92)
at
org.apache.hadoop.hdfs.server.datanode.BlockReceiver.<init>(BlockReceiver.java:174)
at
org.apache.hadoop.hdfs.server.datanode.DataXceiver.writeBlock(DataXceiver.java:454)
at
org.apache.hadoop.hdfs.protocol.datatransfer.Receiver.opWriteBlock(Receiver.java:115)
at
org.apache.hadoop.hdfs.protocol.datatransfer.Receiver.processOp(Receiver.java:68)
at
org.apache.hadoop.hdfs.server.datanode.DataXceiver.run(DataXceiver.java:221)
at java.lang.Thread.run(Thread.java:722)
{quote}
--
This message was sent by Atlassian JIRA
(v6.2#6252)