szucsvillo

> So what is the intention here? We do not use new features of zookeeper and curator, keeping an old version can have most compatibility? > > Thanks. I think...

Example: create table test203 (mykey integer not null primary key, col1 bigint, new_column_1 FLOAT, new_column_2 DOUBLE, new_column_3 BIGINT); create index indexName203 on test203 (col1) INCLUDE (new_column_1,new_column_2,new_column_3); UPSERT INTO TEST203(mykey,col1,new_column_1,new_column_2,new_column_3) VALUES(5,43,201,202,203);...

> IIRC we made use of minikdc from hadoop, has hadoop upgraded kerby? Yes, Hadoop also upgraded kerby, although only from a relatively recent version. https://issues.apache.org/jira/browse/HADOOP-15072

They upgraded to kerby2 https://github.com/apache/hadoop/commit/272844ee572be9de320ce64cc4fc136af6743920