Running MapReduce example from master - fails - need help
se...@blutrumpet.com
What i executed from h2o-samples subproject from h20 (master) March 14, 2014
Class executed:
samples.MapReduce
03:18:45.034 :0 main ----- H2O started -----
03:18:45.045 :0 main Build git branch: (unknown)
03:18:45.045 :0 main Build git hash: (unknown)
03:18:45.045 :0 main Build git describe: (unknown)
03:18:45.046 :0 main Build project version: (unknown)
03:18:45.046 :0 main Built by: '(unknown)'
03:18:45.046 :0 main Built on: '(unknown)'
03:18:45.046 :0 main Java availableProcessors: 8
03:18:45.049 :0 main Java heap totalMemory: 0.12 gb
03:18:45.049 :0 main Java heap maxMemory: 17.78 gb
03:18:45.049 :0 main Java version: Java 1.7.0_51 (from Oracle Corporation)
03:18:45.050 :0 main OS version: Mac OS X 10.9.2 (x86_64)
03:18:45.050 :0 main ICE root: '/tmp/h2o-serge'
03:18:45.054 :0 main Possible IP Address: en0 (en0), fe80:0:0:0:2acf:e9ff:fe1d:ddc7%4
03:18:45.055 :0 main Possible IP Address: en0 (en0), 192.168.1.210
03:18:45.055 :0 main Possible IP Address: lo0 (lo0), fe80:0:0:0:0:0:0:1%1
03:18:45.055 :0 main Possible IP Address: lo0 (lo0), 0:0:0:0:0:0:0:1
03:18:45.055 :0 main Possible IP Address: lo0 (lo0), 127.0.0.1
03:18:45.084 #.168.1.210:54423 main Internal communication uses port: 54424
+ Listening for HTTP and REST traffic on http://192.168.1.210:54423/
03:18:45.085 #.168.1.210:54423 main H2O cloud name: 'serge'
03:18:45.085 #.168.1.210:54423 main (v(unknown)) 'serge' on /192.168.1.210:54423, static configuration based on -flatfile flatfile
03:18:45.085 #.168.1.210:54423 main If you have trouble connecting, try SSH tunneling from your local machine (e.g., via port 55555):
+ 1. Open a terminal and run 'ssh -L 55555:localhost:54423 se...@192.168.1.210'
+ 2. Point your browser to http://localhost:55555
03:18:45.089 #.168.1.210:54423 main Cloud of size 1 formed [/192.168.1.210:54423]
03:18:45.089 #.168.1.210:54423 main Log dir: 'unknown-log-dir'
Exception in thread "main" java.lang.reflect.InvocationTargetException
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:606)
at water.Boot.run(Boot.java:305)
at water.Boot.boot2(Boot.java:281)
at water.Boot.boot(Boot.java:150)
at water.Boot.main(Boot.java:122)
at water.deploy.NodeVM.main(NodeVM.java:19)
Caused by: java.lang.ClassFormatError: StackMapTable format error: bad offset for Uninitialized in method water.api.Inspect2.x1(Lwater/fvec/Vec;JD)Ljava/lang/String;
at water.api.RequestServer.<clinit>(RequestServer.java:134)
at water.H2O.startNetworkServices(H2O.java:1048)
at water.H2O.main(H2O.java:960)
at water.deploy.Cloud$Master.main(Cloud.java:146)
... 9 more
se...@blutrumpet.com
For the record:
Fixed it by commenting out RequestServer.java - line 134
// Request.addToNavbar(registerRequest(new Inspect2()), "Inspect2", "Beta (FluidVecs!)");
Can you guys enable Issues on github please?
I'd up to contribute code following ticket creation.
Thanks
Tom Kraljevic
Hi Serge,
You are welcome to file ticket in our JIRA system. http://jira.0xdata.com
(It's self sign up.)
However, I don't think commenting that out was the real issue.
Probably just some kind of "rebuild project" step wiped out old class files and replaced them with new ones.
java.lang.ClassFormatError suggests some kind of problem by inadvertently mixing incompatible javac or IDE compilers?
Thanks,
Tom
> You received this message because you are subscribed to the Google Groups "H2O Users - h2ostream" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to h2ostream+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
Serge B
Tom Kraljevic
Serge, I (and several folks who develop for H2O) use Intellij, but we don't "create new project" too often, so let me try out that path to make sure it's smooth.
Thanks for pointing out some bumps.
On Mar 14, 2014, at 1:49 PM, Serge B <se...@blutrumpet.com> wrote:
> Thanks Tom for quick reply.
>
> I pulled everything from master earlier, im running IntelliJ
> building (make project) from within it does cause a few errors:
>
> ls[3] = new VecSoftmax(labels, outputStats, NeuralNet.Loss.CrossEntropy);
>
> java: no suitable constructor found for VecSoftmax(water.fvec.Vec,hex.Layer.VecSoftmax,hex.NeuralNet.Loss)
> constructor hex.Layer.VecSoftmax.VecSoftmax(water.fvec.Vec,hex.Layer.VecSoftmax) is not applicable
> (actual and formal argument lists differ in length)
> constructor hex.Layer.VecSoftmax.VecSoftmax() is not applicable
> (actual and formal argument lists differ in length)
>
>
> Changing into:
>
> ls[3] = new VecSoftmax(labels, outputStats);
>
> Fixes the issue since constructor with that other param isnt available.
>
> Could you comment on the other topic i wrote, about deploying mapreduce on a local cloud, something to do with uploading jar and running the job.
The supported models are:
o java -jar h2o.jar on the command line
o running water.Boot inside the IDE will bring up one node inside the IDE
(From inside the IDE)
o java/water/JUnitRunnerDebug.java is a good way to quickly run some example code from inside the IDE.
Thanks,
Tom