Yes it is possible, but the latest Kylin 2.1 version should be used. The previous version has a bug which will write big metadata file for cluster A's HDFS.
2017-09-28 22:03 GMT+08:00 cludezhu <[email protected]>: > "hadoop cluster (A) env will be released " means deleting the cluster. > > after I delete the hadoop cluster, is that possible to keep a kylin query > instance running just with the hbase cluster? > > > thanks > cludezhu > > On September 28, 2017 at 9:27:43 PM, ShaoFeng Shi ([email protected]) > wrote: > > What do you mean by "hadoop cluster (A) env will be released "? Delete the > cluster, or just keep a minimal running? > > If your Hadoop uses on-disk HDFS, there will be data loss when you delete > the cluster. Remember to do a backup for the /kylin folder if you do want > to do that. > > > 2017-09-28 20:12 GMT+08:00 Billy Liu <[email protected]>: > > > sure. That's typical read/write deployment. > > > > 2017-09-28 18:47 GMT+08:00 cludezhu <[email protected]>: > > > > > Hi, > > > > > > currently we have two cluster for kylin deployment: > > > hadoop cluster (A) for calculating and building cube > > > hbase cluster (B) for storing data. > > > > > > We just create the calculating cluster (A) on demand since all our env > is > > > deployed on Cloud. so the hadoop cluster (A) env will be released after > > we > > > finished the calculating job, and only the hbase cluster will be left. > > > > > > our problem is: > > > Is that possible that run a simply kylin query instance on hbase > cluster > > > without the hadoop cluster. > > > > > > any help is appreciate. > > > > > > clude > > > > > > > > > -- > Best regards, > > Shaofeng Shi 史少锋 > -- Best regards, Shaofeng Shi 史少锋
