flume-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rahul Ravindran <rahu...@yahoo.com>
Subject Re: HDFSsink failover error
Date Mon, 14 Jan 2013 22:46:51 GMT
Here is the full config. I swapped the priorities on the sink processor after performing the
namenode failiver and the writes were successful to the newly active namenode.

agent1.channels.ch1.type = FILE
agent1.channels.ch1.checkpointDir = /flume_runtime/checkpoint
agent1.channels.ch1.dataDirs = /flume_runtime/data


agent1.channels.ch2.type = FILE
agent1.channels.ch2.checkpointDir = /flume_runtime/checkpoint2
agent1.channels.ch2.dataDirs = /flume_runtime/data2



# Define an Avro source called avro-source1 on agent1 and tell it

# to bind to 0.0.0.0:41414. Connect it to channel ch1.

agent1.sources.avro-source1.channels = ch1
agent1.sources.avro-source1.type = avro
agent1.sources.avro-source1.bind = 0.0.0.0
agent1.sources.avro-source1.port = 4545



agent1.sources.avro-source2.channels = ch2
agent1.sources.avro-source2.type = avro
agent1.sources.avro-source2.bind = 0.0.0.0
agent1.sources.avro-source2.port = 4546


agent1.sinks.hdfs-sink1.channel = ch1
agent1.sinks.hdfs-sink1.type = hdfs
agent1.sinks.hdfs-sink1.hdfs.path = hdfs://ip-10-4-71-187.ec2.internal/user/br/shim/eventstream/event/host101/
agent1.sinks.hdfs-sink1.hdfs.filePrefix = event
agent1.sinks.hdfs-sink1.hdfs.writeFormat = Text
agent1.sinks.hdfs-sink1.hdfs.rollInterval = 120
agent1.sinks.hdfs-sink1.hdfs.rollCount = 0
agent1.sinks.hdfs-sink1.hdfs.rollSize = 0
agent1.sinks.hdfs-sink1.hdfs.fileType = DataStream
agent1.sinks.hdfs-sink1.hdfs.batchSize = 1000
agent1.sinks.hdfs-sink1.hdfs.txnEventSize = 1000

agent1.sinks.hdfs-sink2.channel = ch2
agent1.sinks.hdfs-sink2.type = hdfs
agent1.sinks.hdfs-sink2.hdfs.path = hdfs://ip-10-4-71-187.ec2.internal/user/br/shim/eventstream/event/host102/
agent1.sinks.hdfs-sink2.hdfs.filePrefix = event
agent1.sinks.hdfs-sink2.hdfs.writeFormat = Text
agent1.sinks.hdfs-sink2.hdfs.rollInterval = 120
agent1.sinks.hdfs-sink2.hdfs.rollCount = 0
agent1.sinks.hdfs-sink2.hdfs.rollSize = 0
agent1.sinks.hdfs-sink2.hdfs.fileType = DataStream
agent1.sinks.hdfs-sink2.hdfs.batchSize = 1000
agent1.sinks.hdfs-sink2.hdfs.txnEventSize = 1000


agent1.sinks.hdfs-sink1-back.channel = ch1
agent1.sinks.hdfs-sink1-back.type = hdfs
agent1.sinks.hdfs-sink1-back.hdfs.path = hdfs://ip-10-110-69-240.ec2.internal/user/br/shim/eventstream/event/host101/
agent1.sinks.hdfs-sink1-back.hdfs.filePrefix = event
agent1.sinks.hdfs-sink1-back.hdfs.writeFormat = Text
agent1.sinks.hdfs-sink1-back.hdfs.rollInterval = 120
agent1.sinks.hdfs-sink1-back.hdfs.rollCount = 0
agent1.sinks.hdfs-sink1-back.hdfs.rollSize = 0
agent1.sinks.hdfs-sink1-back.hdfs.fileType = DataStream
agent1.sinks.hdfs-sink1-back.hdfs.batchSize = 1000
agent1.sinks.hdfs-sink1-back.hdfs.txnEventSize = 1000

agent1.sinks.hdfs-sink2-back.channel = ch2
agent1.sinks.hdfs-sink2-back.type = hdfs
agent1.sinks.hdfs-sink2-back.hdfs.path = hdfs://ip-10-110-69-240.ec2.internal/user/br/shim/eventstream/event/host102/
agent1.sinks.hdfs-sink2-back.hdfs.filePrefix = event
agent1.sinks.hdfs-sink2-back.hdfs.writeFormat = Text
agent1.sinks.hdfs-sink2-back.hdfs.rollInterval = 120
agent1.sinks.hdfs-sink2-back.hdfs.rollCount = 0
agent1.sinks.hdfs-sink2-back.hdfs.rollSize = 0
agent1.sinks.hdfs-sink2-back.hdfs.fileType = DataStream
agent1.sinks.hdfs-sink2-back.hdfs.batchSize = 1000
agent1.sinks.hdfs-sink2-back.hdfs.txnEventSize = 1000



agent1.sinkgroups.failoverGroup1.sinks = hdfs-sink1 hdfs-sink1-back
agent1.sinkgroups.failoverGroup1.processor.type = failover
#higher number in priority is higher priority
agent1.sinkgroups.failoverGroup1.processor.priority.hdfs-sink1 = 10
agent1.sinkgroups.failoverGroup1.processor.priority.hdfs-sink1-back = 5
#failover if failure detected for 10 seconds
agent1.sinkgroups.failoverGroup1.processor.maxpenalty = 10000


agent1.sinkgroups.failoverGroup2.sinks = hdfs-sink2 hdfs-sink2-back
agent1.sinkgroups.failoverGroup2.processor.type = failover
#higher number in priority is higher priority
agent1.sinkgroups.failoverGroup2.processor.priority.hdfs-sink2 = 10
agent1.sinkgroups.failoverGroup2.processor.priority.hdfs-sink2-back = 5
#failover if failure detected for 10 seconds
agent1.sinkgroups.failoverGroup2.processor.maxpenalty = 10000

# Finally, now that we've defined all of our components, tell
# agent1 which ones we want to activate.
agent1.sinkgroups = failoverGroup1 failoverGroup2
agent1.channels = ch1 ch2
agent1.sources = avro-source1 avro-source2
agent1.sinks = hdfs-sink1 hdfs-sink2 hdfs-sink1-back hdfs-sink2-back


________________________________
 From: Connor Woodson <cwoodson.dev@gmail.com>
To: user@flume.apache.org; Rahul Ravindran <rahulrv@yahoo.com> 
Sent: Monday, January 14, 2013 2:28 PM
Subject: Re: HDFSsink failover error
 

I assume that's only part of your config as it's missing a source; if you get rid of the sink
processor, can you write to each hdfs sink individually? (comment one out at a time)

- Connor



On Mon, Jan 14, 2013 at 1:42 PM, Rahul Ravindran <rahulrv@yahoo.com> wrote:

Hi,
>   I am attempting to setup an HDFS sink such that when a namenode failover occurs (active
namenode is brought down and the standby namenode switches to active), the failover sink would
send events to the new active namenode. I see an error about WRITE not supported in standby
state..Does this not count as a failure for the failover sink?
>Thanks,
>~Rahul.
>
>
>My config is as follows:
>
>
>agent1.sinks.hdfs-sink1.channel = ch1
>agent1.sinks.hdfs-sink1.type = hdfs
>agent1.sinks.hdfs-sink1.hdfs.path = hdfs://ip-10-4-71-187.ec2.internal/user/br/shim/eventstream/event/host101/
>agent1.sinks.hdfs-sink1.hdfs.filePrefix = event
>agent1.sinks.hdfs-sink1.hdfs.writeFormat = Text
>agent1.sinks.hdfs-sink1.hdfs.rollInterval = 120
>agent1.sinks.hdfs-sink1.hdfs.rollCount = 0
>agent1.sinks.hdfs-sink1.hdfs.rollSize = 0
>agent1.sinks.hdfs-sink1.hdfs.fileType = DataStream
>agent1.sinks.hdfs-sink1.hdfs.batchSize = 1000
>agent1.sinks.hdfs-sink1.hdfs.txnEventSize = 1000
>
>
>agent1.sinks.hdfs-sink1-back.channel = ch1
>agent1.sinks.hdfs-sink1-back.type = hdfs
>agent1.sinks.hdfs-sink1-back.hdfs.path = hdfs://ip-10-110-69-240.ec2.internal/user/br/shim/eventstream/event/host101/
>agent1.sinks.hdfs-sink1-back.hdfs.filePrefix = event
>agent1.sinks.hdfs-sink1-back.hdfs.writeFormat = Text
>agent1.sinks.hdfs-sink1-back.hdfs.rollInterval = 120
>agent1.sinks.hdfs-sink1-back.hdfs.rollCount = 0
>agent1.sinks.hdfs-sink1-back.hdfs.rollSize = 0
>agent1.sinks.hdfs-sink1-back.hdfs.fileType = DataStream
>agent1.sinks.hdfs-sink1-back.hdfs.batchSize = 1000
>agent1.sinks.hdfs-sink1-back.hdfs.txnEventSize = 1000
>
>
>agent1.sinkgroups.failoverGroup1.sinks = hdfs-sink1 hdfs-sink1-back
>agent1.sinkgroups.failoverGroup1.processor.type = failover
>#higher number in priority is higher priority
>agent1.sinkgroups.failoverGroup1.processor.priority.hdfs-sink1 = 10
>agent1.sinkgroups.failoverGroup1.processor.priority.hdfs-sink1-back = 5
>#failover if failure detected for 10 seconds
>agent1.sinkgroups.failoverGroup1.processor.maxpenalty = 10000
>
>
>
>
>agent1.sinkgroups = failoverGroup1
>
>
>
>14 Jan 2013 21:37:28,819 INFO  [hdfs-hdfs-sink2-call-runner-6] (org.apache.flume.sink.hdfs.BucketWriter.doOpen:208)
 - Creating hdfs://ip-10-4-71-187.ec2.internal/....
>14 Jan 2013 21:37:28,834 WARN  [SinkRunner-PollingRunner-FailoverSinkProcessor] (org.apache.flume.sink.hdfs.HDFSEventSink.process:456)
 - HDFS IO error
>org.apache.hadoop.ipc.RemoteException(org.apache.hadoop.ipc.StandbyException): Operation
category WRITE is not supported in state standby
>        at org.apache.hadoop.hdfs.server.namenode.ha.StandbyState.checkOperation(StandbyState.java:87)
>        at org.apache.hadoop.hdfs.server.namenode.NameNode$NameNodeHAContext.checkOperation(NameNode.java:1379)
>        at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.checkOperation(FSNamesystem.java:762)
>        at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.startFileInt(FSNamesystem.java:1688)
>        at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.startFile(FSNamesystem.java:1669)
>        at org.apache.hadoop.hdfs.server.namenode.NameNodeRpcServer.create(NameNodeRpcServer.java:409)
>        at org.apache.hadoop.hdfs.protocolPB.ClientNamenodeProtocolServerSideTranslatorPB.create(ClientNamenodeProtocolServerSideTranslatorPB.java:205)
>        at org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos$ClientNamenodeProtocol$2.callBlockingMethod(ClientNamenodeProtocolProtos.java:44068)
>        at org.apache.hadoop.ipc.ProtobufRpcEngine$Server$ProtoBufRpcInvoker.call(ProtobufRpcEngine.java:453)
>        at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:898)
>        at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:1693)
>        at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:1689)
>        at java.security.AccessController.doPrivileged(Native Method)
>        at javax.security.auth.Subject.doAs(Subject.java:396)
>        at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1332)
>        at org.apache.hadoop.ipc.Server$Handler.run(Server.java:1687)
>
>
>        at org.apache.hadoop.ipc.Client.call(Client.java:1160)
>        at org.apache.hadoop.ipc.ProtobufRpcEngine$Invoker.invoke(ProtobufRpcEngine.java:202)
>        at $Proxy11.create(Unknown Source)
>        at sun.reflect.GeneratedMethodAccessor8.invoke(Unknown Source)
>        at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
>        at java.lang.reflect.Method.invoke(Method.java:616)
>        at org.apache.hadoop.io.retry.RetryInvocationHandler.invokeMetho
Mime
View raw message