[ https://issues.apache.org/jira/browse/HDFS-2756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13180993#comment-13180993
]
Hari Mankude commented on HDFS-2756:
------------------------------------
Standby is in safe mode. Does not read the in_progress edit log. This could result in lots
of edit log entries that need to be applied at the standby before it can start functioning
as a active NN during failover.
2012-01-06 00:32:06,600 INFO ipc.Server (Server.java:run(712)) - IPC Server Responder: starting
2012-01-06 00:32:06,601 INFO ipc.Server (Server.java:run(543)) - IPC Server listener on 9000:
starting
2012-01-06 00:32:06,601 INFO namenode.NameNode (NameNode.java:startCommonServices(444)) -
NameNode up at: hrt8n17.cc1.ygridcore.net/98.137.233.237:9000
2012-01-06 00:32:06,602 INFO namenode.FSNamesystem (FSNamesystem.java:startStandbyServices(577))
- Starting services required for standby state
2012-01-06 00:32:06,603 INFO ha.EditLogTailer (EditLogTailer.java:doTailEdits(111)) - lastTxnId:
2974
2012-01-06 00:32:06,610 INFO ha.StandbyCheckpointer (StandbyCheckpointer.java:start(126))
- Starting standby checkpoint thread...
Checkpointing active NN at 0.0.0.0:50070
Serving checkpoints at /0.0.0.0:50070
2012-01-06 00:32:06,790 INFO ha.EditLogTailer (EditLogTailer.java:doTailEdits(116)) - edit
streams to load from: 0
2012-01-06 00:32:06,790 INFO ha.EditLogTailer (EditLogTailer.java:doTailEdits(119)) - editsLoaded:
0
2012-01-06 00:32:07,102 INFO hdfs.StateChange (DatanodeManager.java:registerDatanode(573))
- BLOCK* NameSystem.registerDatanode: node registration from 98.137.233.237:50010 storage
DS-888494204-98.137.233.237-50010-1325801757323
2012-01-06 00:32:07,105 INFO net.NetworkTopology (NetworkTopology.java:add(344)) - Adding
a new node: /default-rack/98.137.233.237:50010
2012-01-06 00:32:07,106 INFO namenode.FSNamesystem (FSNamesystem.java:checkMode(3411)) -
DID NOT call initialize at 1
2012-01-06 00:32:07,106 INFO hdfs.StateChange (DatanodeManager.java:registerDatanode(573))
- BLOCK* NameSystem.registerDatanode: node registration from 98.137.233.235:50010 storage
DS-1976177484-98.137.233.235-50010-1325801709842
2012-01-06 00:32:07,108 INFO net.NetworkTopology (NetworkTopology.java:add(344)) - Adding
a new node: /default-rack/98.137.233.235:50010
2012-01-06 00:32:07,111 INFO namenode.FSNamesystem (FSNamesystem.java:checkMode(3411)) -
DID NOT call initialize at 1
2012-01-06 00:32:07,121 INFO namenode.NameNode (NameNodeRpcServer.java:blockReport(861))
- Required GS=2459 BlockReport GS 3059, Queuing blockReport message
2012-01-06 00:32:07,156 INFO namenode.NameNode (NameNodeRpcServer.java:blockReport(861))
- Required GS=2459 BlockReport GS 3059, Queuing blockReport message
> Warm standby does not read the in_progress edit log
> ----------------------------------------------------
>
> Key: HDFS-2756
> URL: https://issues.apache.org/jira/browse/HDFS-2756
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: ha
> Affects Versions: HA branch (HDFS-1623)
> Reporter: Hari Mankude
>
> Warm standby does not read the in_progress edit log. This could result in standby taking
a long time to become the primary during a failover scenario.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
|