If you use Ambari to install the cluster, you can add a client node for running Kylin;
For other cases, you can add a clean machine for Kylin, and then manually copy the Hadoop/hive/hbase clients from a Hadoop node to the Kylin server. If Hive/hbase command line can work from it, then Kylin should be able to start. 2017-12-22 13:39 GMT+08:00 Qiwen.li <[email protected]>: > you may choose to use a sandbox which is a standalone node. > As for how to configure kylin on a sandbox, you can refrence the article: > http://kylin.apache.org/docs21/install/hadoop_env.html > > > > > 发自网易邮箱大师 > > > On 12/22/2017 13:34，张 佑铖<[email protected]> wrote： > Hi kylin team: > > > I want to deploy kylin in a standalone node apart from the hadoop cluster. > > Can you give me some hints or successful cases? > > > As I know, when run kylin.sh start, the script will call 'hbase classpath' > and `hive -e set | grep 'env:CLASSPATH'`, how can I find a workaround ? Is > there any other method? > > Is it possible to use client tool instead of shell like thrift, jdbc or > something else? > > > > Thank you > > youcheng > > > -- Best regards, Shaofeng Shi 史少锋
