0

I'm trying to start a H2O cluster as external backend for Sparkling Water manually. By following the documentation here it says I need to use the parameter 'name' with the extended H2O driver. But by doing so it says that the parameter 'name' doesn't exist. How can I set the cloud name so that I can use it to identify the cluster in the sparkling-shell? The help of the driver also doesn't mention a way to set the cloud name.

Any help would be appreciated.

Best regards,

Markus

Markus Wilhelm
  • 171
  • 2
  • 2
  • 11

2 Answers2

0

Try this one:

java -cp $H2O_EXTENDED_JAR water.H2OApp -name test

It looks like a bug in documentation.

Michal
  • 437
  • 3
  • 8
0

The issue is that when you have extended h2o jar for h2odriver (with support for particular Hadoop), this h2odriver truly does not have -name argument. Instead, you need to use -jobname argument instead. We need to fix our documentation for this case.

If you are using the extended h2o driver for standalone h2o application, the -name argument should do the job.

Jakub Háva
  • 228
  • 1
  • 5