Hi ChengCheng
Here is an old discussion about migration, take a look and hope it helps: http://apache-kylin.74782.x6.nabble.com/How-to-migrate-cube-from-one-Kylin-to-another-for-example-Kylin-QA-to-Kylin-Production-within-the-dir-td8628.html And you can read this FAQ, it also helps: http://kylin.apache.org/docs/gettingstarted/faq.html A1: For metadata, back up and restore is pretty simple, see this: http://kylin.apache.org/docs/howto/howto_backup_metadata.html. And for the others, you may choose your own solution after you go through the discussion. Here I quote Yang's solution: "Rebuild the cube in new cluster is often easier" A2: You can find your HDFS data dir in kylin.properties: "kylin.env.hdfs-working-dir", default is "/kylin", more explanation see in FAQ. A3: You can run tools to clean up your metadata and your storage data, see this: http://kylin.apache.org/docs/howto/howto_cleanup_storage.html. --------- Regards! Aron Tao ChengCheng Liu <[email protected]> 于2018年10月27日周六 上午6:47写道： > Hi, guys > > As we all known, kylin produce two kinds of hdfs data. One kind data > is produced by kylin server, and the other one is produced by hbase. Kylin > has run one year. It is required to switch hdfs cluster. > > Q1: Do both hdfs data for hbase and dfs data for kylin server have > to be migrated together? > > Q2: Where is the hdfs data dir for hbase? Where is the hdfs data dir > for kylin server? > > Q3: I find hdfs working dir exists lots of data file, how to delete > useless file? > >
