[
https://issues.apache.org/jira/browse/HDFS-3924?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13454434#comment-13454434
]
Andy Isaacson commented on HDFS-3924:
-------------------------------------
{code}
@@ -234,8 +235,8 @@ class BlockStorageLocationUtil {
}
// Get the VolumeId by indexing into the list of VolumeIds
// provided by the datanode
- HdfsVolumeId id = new HdfsVolumeId(metaVolumeIds.get(volumeIndex)[0],
- true);
+ int volumeId =
ByteBuffer.wrap(metaVolumeIds.get(volumeIndex)).getInt();
+ HdfsVolumeId id = new HdfsVolumeId(volumeId, true);
// Find out which index we are in the LocatedBlock's replicas
LocatedBlock locBlock = extBlockToLocBlock.get(extBlock);
DatanodeInfo[] dnInfos = locBlock.getLocations();
{code}
This change is a wire compatibility change; the old code puts 1 byte in the
protobuf and the new code requires 4 bytes to decode (else {{.getInt}} throws
BufferUnderflowException), unless I'm missing something.
Maybe we don't care about wire compatibility, but if so we should explicitly
state that we don't care.
> Multi-byte id in HdfsVolumeId
> -----------------------------
>
> Key: HDFS-3924
> URL: https://issues.apache.org/jira/browse/HDFS-3924
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs client
> Affects Versions: 2.0.3-alpha
> Reporter: Andrew Wang
> Assignee: Andrew Wang
> Attachments: hdfs-3924-1.patch
>
>
> Currently, the {{id}} in {{HdfsVolumeId}} is just a single byte. It'd be nice
> to make this larger, for situations where there's a large number of disks.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
