Many is asking or have hit problem running Kylin on HBase 1.x or HDP 2.3. So I spend some time testing on a HDP 2.3.1 sandbox and here is the status.
Now the *1.x-HBase1.1.3* branch contains latest code that can work on HBase 1.1.3 or above. Known problems like fail to find hive dependency and sample cube not working have all been fixed. All unit test and integration test passed on the 2.3.1 sandbox (except one thing, more on this below), sample cube can build and query. I will maintain *1.x-HBase1.1.3* regularly by taking the latest 1.x-staging and apply a patch on top. Note the requirement of HBase version 1.1.3 (or above). There is a known bug in HBase earlier versions about fuzzy key filter that will cause missing rows or lesser aggregations in Kylin result. https://issues.apache.org/jira/browse/HBASE-14269 Cheers Yang
