Ming Ma created HDFS-6943:
-----------------------------
Summary: Improve NN allocateBlock log to include replicas'
datanode IPs
Key: HDFS-6943
URL: https://issues.apache.org/jira/browse/HDFS-6943
Project: Hadoop HDFS
Issue Type: Improvement
Reporter: Ming Ma
Datanode storage ID used to use IP and port. It has changed to use UUID. This
makes debugging harder when we want to understand which DNs are assigned when
DFSClient calls addBlock. For example,
{noformat}
BLOCK* allocateBlock: /foo. BP-1980237412-xx.xx.xxx.xxx-1408142057773
blk_1227779764_154043834{blockUCState=UNDER_CONSTRUCTION, primaryNodeIndex=-1,
replicas=[ReplicaUnderConstruction[[DISK]DS-9479727b-24c5-4068-8703-dfb9a41c056c:NORMAL|RBW],
ReplicaUnderConstruction[[DISK]DS-abe7840c-1db8-4623-9da7-3aed6a28c4f4:NORMAL|RBW],
ReplicaUnderConstruction[[DISK]DS-956023f4-56a0-4c30-a148-b78c61cf764b:NORMAL|RBW]]}
{noformat}
--
This message was sent by Atlassian JIRA
(v6.2#6252)
