Personally from my experience having your hive metastore on Oracle offers the best concurrency. Oracle locking and concurrency is best suited for metastore to avoid deadlocks and excessive serialisation.
On 26/8/2015, "Subroto Sanyal" <ssan...@datameer.com> wrote: >hi Jerrick > >Yes, thats the correct way to interact with HiveMetastore(provided you are >only interested to play around metadata and Hive Metastore Service is >accessible from the java client). >Further, you can try the Hive JDBC driver or Hive Server2 thrift API >(couldn't find the documentation for this but, it works for me) to access >the metadata and physical data. > >On Wed, Aug 26, 2015 at 9:07 AM, Dr Mich Talebzadeh <m...@peridale.co.uk> >wrote: > >> >> What is your hive metastore is on? >> >> If it is on an RDBMS, then you can simply use client/JDBC tools to access >> the metastore. Pretty simple. >> >> HTH, >> >> On 26/8/2015, "Jerrick Hoang" <jerrickho...@gmail.com> wrote: >> >> >Hi all, >> > >> >I want to interact with HiveMetaStore table from code and was looking at >> > >> http://hive.apache.org/javadocs/r0.13.1/api/metastore/org/apache/hadoop/hive/metastore/HiveMetaStoreClient.html >> >, was wondering if this is the correct way to do this or should I use a >> >jdbc client. If HiveMetaStoreClient is the right way then how do I create >> a >> >client given a hive-site.xml, >> > >> >Thanks, >> >J >> > >> > >> > >
