Hi, Edward,
Thank you.
I updated the version. There comes a new problem.
In the /Apache Hama BSP Programming Model
(http://people.apache.org/~tjungblut/downloads/hamadocs/ApacheHamaBSPProgrammingmodel_06.pdf)/
there is a paragraph describes 'Graph repair' : "Hama requires a graph
to be completed before feeding it to an algorithm. By complete we mean
that every vertex that is referenced by an edge must somewhere be a
vertex in the graph. In many cases of leafs this is not always the case,
therefore we have added a repair functionality which is traversing the
whole graph for leafs and adding them to the vertex structure to prevent
algorithms from breaking with NullPointerExceptions when it does not
find a referenced vertex. You can turn this feature on by setting it in
your configuration like this:
conf.setBoolean(GraphJobRunner.GRAPH_REPAIR, true);"
I followed the guidance but got a hint 'GRAPH_REPAIR cannot be resolved
or is not a field' programming in Eclipse. I read the source code later
and there is really no variable named GRAPH_REPAIR in
GraphJobRunner.java which exists in the version 0.5.0. Is this function
not supported any longer ? How can I make a 'repaired graph'?
Thank you.
On 04/17/2013 12:51 PM, Edward J. Yoon wrote:
Please use 0.6.1 and try your application with small data again.
See also http://hama.apache.org/run_examples.html
On Wed, Apr 17, 2013 at 10:27 AM, Lyu Xuedong <[email protected]> wrote:
hama: 0.6.0
hadoop : 1.0.4
JDK : 1.6
OS : ubuntu 12.04
On 04/17/2013 05:47 AM, Edward J. Yoon wrote:
Your version?
On Wed, Apr 17, 2013 at 12:07 AM, Lvxuedong <[email protected]> wrote:
Hi, Edward, thank you. But your suggestion seems do not work, do you have
some other advice ?
Is java.lang.NullPointerException related with heap size ?
在 2013-4-16，22:13，"Edward J. Yoon" <[email protected]> 写道：
I guess you need to increase the child processor JVM heap size.
- conf/hama-site.xml:
<property>
<name>bsp.child.java.opts</name>
<value>-Xmx2048m</value>
</property>
On Tue, Apr 16, 2013 at 10:55 PM, Lyu Xuedong <[email protected]>
wrote:
My project can work well in eclipse, but when I export it as a jar file
and
submit it to a two-node hama cluster, some errors and fatal occur if
the
input file is over 64M.
I suspect that my configure files of hadoop or hama is somewhere not
right,
but a pi estimator can run normally on my cluster and
GroomServer$BSPPerChild can be seen on each nodes. I debug for a whole
day,
nothing improved.
What in my input file are a large number of RDF triples: "<subject>
<predicate> <object> ." My task is to create vertices for subjects and
objects. Predicates are subjects' edges.
What should I do ?
Terminal output:
13/04/16 21:13:36 INFO bgp.HamaBgpComplete: Job begain.
13/04/16 21:13:37 INFO bsp.FileInputFormat: Total input paths to
process : 2
13/04/16 21:13:38 INFO bsp.BSPJobClient: Running job:
job_201304161357_0015
13/04/16 21:13:41 INFO bsp.BSPJobClient: Current supersteps number: 0
13/04/16 21:13:47 INFO bsp.BSPJobClient: Current supersteps number: 2
13/04/16 21:13:53 INFO bsp.BSPJobClient: Current supersteps number: 3
13/04/16 21:13:59 INFO bsp.BSPJobClient: Current supersteps number: 4
13/04/16 21:14:05 INFO bsp.BSPJobClient: Current supersteps number: 5
attempt_201304161357_0015_000000_0: 13/04/16 21:13:43 INFO
sync.ZKSyncClient: Initializing ZK Sync Client
attempt_201304161357_0015_000000_0: 13/04/16 21:13:43 INFO
sync.ZooKeeperSyncClientImpl: Start connecting to Zookeeper! At
hadoop1/1.2.3.4:61002
attempt_201304161357_0015_000000_0: 13/04/16 21:13:43 INFO ipc.Server:
Starting SocketReader
attempt_201304161357_0015_000000_0: 13/04/16 21:13:43 INFO ipc.Server:
IPC
Server Responder: starting
attempt_201304161357_0015_000000_0: 13/04/16 21:13:43 INFO ipc.Server:
IPC
Server handler 0 on 61002: starting
attempt_201304161357_0015_000000_0: 13/04/16 21:13:43 INFO
message.HadoopMessageManagerImpl: BSPPeer address:hadoop1 port:61002
attempt_201304161357_0015_000000_0: 13/04/16 21:13:43 INFO ipc.Server:
IPC
Server listener on 61002: starting
attempt_201304161357_0015_000000_0: 13/04/16 21:14:03 ERROR
bsp.BSPTask:
Error running bsp setup and bsp function.
attempt_201304161357_0015_000000_0: java.lang.NullPointerException
attempt_201304161357_0015_000000_0: 13/04/16 21:14:04 INFO ipc.Server:
Stopping server on 61002
attempt_201304161357_0015_000000_0: 13/04/16 21:14:04 INFO ipc.Server:
IPC
Server handler 0 on 61002: exiting
attempt_201304161357_0015_000000_0: 13/04/16 21:14:04 INFO ipc.Server:
Stopping IPC Server listener on 61002
attempt_201304161357_0015_000000_0: 13/04/16 21:14:04 INFO ipc.Server:
Stopping IPC Server Responder
attempt_201304161357_0015_000000_0: 13/04/16 21:14:04 INFO
metrics.RpcInstrumentation: shut down
attempt_201304161357_0015_000000_0: 13/04/16 21:14:04 ERROR
bsp.BSPTask:
Shutting down ping service.
attempt_201304161357_0015_000000_0: 13/04/16 21:14:04 FATAL
bsp.GroomServer:
Error running child
attempt_201304161357_0015_000000_0: java.lang.NullPointerException
attempt_201304161357_0015_000000_0: java.lang.NullPointerException
13/04/16 21:15:11 INFO bsp.BSPJobClient: Job failed.
tasklogs:
attempt_201304161357_0015_000000_0.log
13/04/16 21:13:43 INFO sync.ZKSyncClient: Initializing ZK Sync Client
13/04/16 21:13:43 INFO sync.ZooKeeperSyncClientImpl: Start connecting
to
Zookeeper! At hadoop1/1.2.3.4:61002
13/04/16 21:13:43 INFO ipc.Server: Starting SocketReader
13/04/16 21:13:43 INFO ipc.Server: IPC Server Responder: starting
13/04/16 21:13:43 INFO ipc.Server: IPC Server handler 0 on 61002:
starting
13/04/16 21:13:43 INFO message.HadoopMessageManagerImpl: BSPPeer
address:hadoop1 port:61002
13/04/16 21:13:43 INFO ipc.Server: IPC Server listener on 61002:
starting
13/04/16 21:14:03 ERROR bsp.BSPTask: Error running bsp setup and bsp
function.
java.lang.NullPointerException
13/04/16 21:14:04 INFO ipc.Server: Stopping server on 61002
13/04/16 21:14:04 INFO ipc.Server: IPC Server handler 0 on 61002:
exiting
13/04/16 21:14:04 INFO ipc.Server: Stopping IPC Server listener on
61002
13/04/16 21:14:04 INFO ipc.Server: Stopping IPC Server Responder
13/04/16 21:14:04 INFO metrics.RpcInstrumentation: shut down
13/04/16 21:14:04 ERROR bsp.BSPTask: Shutting down ping service.
13/04/16 21:14:04 FATAL bsp.GroomServer: Error running child
java.lang.NullPointerException
java.lang.NullPointerException
attempt_201304161357_0015_000001_0.log
13/04/16 21:13:42 INFO sync.ZKSyncClient: Initializing ZK Sync Client
13/04/16 21:13:42 INFO sync.ZooKeeperSyncClientImpl: Start connecting
to
Zookeeper! At hadoop1/1.2.3.4:61001
13/04/16 21:13:42 ERROR sync.ZooKeeperSyncClientImpl:
org.apache.zookeeper.KeeperException$NoNodeException: KeeperErrorCode =
NoNode for /bsp/job_201304161357_0015/peers
13/04/16 21:13:42 INFO ipc.Server: Starting SocketReader
13/04/16 21:13:42 INFO ipc.Server: IPC Server Responder: starting
13/04/16 21:13:42 INFO message.HadoopMessageManagerImpl: BSPPeer
address:hadoop1 port:61001
13/04/16 21:13:42 INFO ipc.Server: IPC Server listener on 61001:
starting
13/04/16 21:13:42 INFO ipc.Server: IPC Server handler 0 on 61001:
starting
13/04/16 21:14:06 ERROR bsp.BSPPeerImpl: Error while sending messages
java.io.IOException: Call to hadoop1/1.2.3.4:61002 failed on local
exception: java.io.EOFException
at org.apache.hadoop.ipc.Client.wrapException(Client.java:1103)
at org.apache.hadoop.ipc.Client.call(Client.java:1071)
at org.apache.hadoop.ipc.RPC$Invoker.invoke(RPC.java:225)
at $Proxy3.put(Unknown Source)
at
org.apache.hama.bsp.message.HadoopMessageManagerImpl.transfer(HadoopMessageManagerImpl.java:108)
at org.apache.hama.bsp.BSPPeerImpl.sync(BSPPeerImpl.java:410)
at
org.apache.hama.graph.GraphJobRunner.bsp(GraphJobRunner.java:118)
at org.apache.hama.bsp.BSPTask.runBSP(BSPTask.java:166)
at org.apache.hama.bsp.BSPTask.run(BSPTask.java:143)
at
org.apache.hama.bsp.GroomServer$BSPPeerChild.main(GroomServer.java:1266)
Caused by: java.io.EOFException
at java.io.DataInputStream.readInt(DataInputStream.java:375)
at
org.apache.hadoop.ipc.Client$Connection.receiveResponse(Client.java:800)
at org.apache.hadoop.ipc.Client$Connection.run(Client.java:745)
--
Best Regards, Edward J. Yoon
@eddieyoon
