You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@predictionio.apache.org by 赵睿 <zo...@163.com> on 2016/08/31 02:47:58 UTC
Integrate PredictionIO into cluster
Hi,
I have being developed with PredictionIO in standalone mode for almost one years, everything worked well.
Recently, I tried to integrate PredictionIO into cluster, including spark、hbase and hdfs.
It was fine to install and run predictionIO platform, but I met a difficult problem when I tried to run pio-template-fpm-master.
The problem occurs when "pio train" is launch, error is below:
This error occurs when pio fetch data from hbase. it's OK to save data into hbase. I guess it's a linking problem, but I don't know how to fix it?
There are so little information about how to integrate integrate pio into cluster. I just change pio_env.sh depending on my experience.
In My environment:
cluster is installed through cloudera-5.7.1, with hbase-1.2.0 and spark-1.6.0
pio's version is 0.9.5, which compiling with hbase-0.98.5
hbase is used for pio_event.
My questions are:
Can pio run very well in cluster?
--
best regards
赵睿
Tel:13466571475