On Tue, Mar 25, 2014 at 6:56 PM, George Chow <[email protected]> wrote:
> Hi Venki, > > I made a mistake initially and thought I was embedded. But I'm actually in > scenario 1 (remote metastore). > I am setting hive.metastore.uris to my HiveServer's address and port. Is > that correct? > hive.metastore.uris should point to metastore server which is different from HiveServer1/2. Drill uses Metastore thrift interface to fetch the table/db info. From your first email, it looks like you are embedding the metastore within HiveServer1. You can copy the following settings from you current hiveserver's hive-site.xml to storage-engines.json to embed metastore in Drillbit or start the metastore as separate service using 'hive --service metastore' and point hive.metastore.uris to the address. "javax.jdo.option.ConnectionURL" : "jdbc:mysql://hostname/hivedrill?createDatabaseIfNotExist=true", "javax.jdo.option.ConnectionDriverName" : "com.mysql.jdbc.Driver", "javax.jdo.option.ConnectionUserName" : "user", "javax.jdo.option.ConnectionPassword" : "passwd" Thanks Venki > George > > > On Tue, Mar 25, 2014 at 6:48 PM, Venki Korukanti > <[email protected]>wrote: > > > Hi George, > > > > Are you embedding the metastore in drillbit (hive.metastore.uris="") or > > drillbit is connecting to remote metastore (hive.metastore.uris=valid > > address)? > > > > If you are embedding the metastore, make sure you have the appropriate > > driver (if MySql backend then mysql jdbc driver) in classpath or > drill/lib > > folder. > > > > Thanks > > Venki > > > > > > On Tue, Mar 25, 2014 at 6:30 PM, George Chow <[email protected]> > wrote: > > > > > Hi, > > > > > > Thanks Tim for the tip. It helps... so I copied the settings from my > > > hive-site.xml over to storage-engines.json. > > > > > > I restart my drillbit but I still don't see the contents from my Hive > > > metastore. The log file (/var/log/drill/node11/drillbits.out) doesn't > > > contain anything about the inclusion of Hive. > > > > > > (I'm assuming that SELECT * FROM INFORMATION_SCHEMA.TABLES will return > > more > > > than the usual 5 tables (VIEW/COLUMNS/TABLES/CATALOGS/SCHEMATA.) > > > > > > George > > > > > > > > > On Tue, Mar 25, 2014 at 6:12 PM, Timothy Chen <[email protected]> > wrote: > > > > > > > Hi George, > > > > > > > > I believe we have a wiki section on setting up hive + drill : > > > > > > > > > https://cwiki.apache.org/confluence/display/DRILL/Hive+Storage+Engine > > > > > > > > I haven't personally tried it yet but see if this helps. > > > > > > > > Tim > > > > > > > > > > > > On Tue, Mar 25, 2014 at 6:10 PM, George Chow <[email protected]> > > > wrote: > > > > > > > > > Hi, > > > > > > > > > > So I'm trying to setup a drillbit to reach across to my Hive > > > metastore. I > > > > > dug into JIRA and see that DRILL-354/357 are both resolved. I've > also > > > > > looked into my drillbit and found .../conf/storage-engines.json. > > > > > > > > > > Inside, I see an entry for Hive that's commented out. However, it > is > > > for > > > > an > > > > > instance Hive with a Derby metastore. My instance is a HiveServer1 > > > with a > > > > > MySQL metastore. So both javax.jdo.option.ConnectionURL and > > > > > hive.metastore.sasl.enabled are clear to me. But what about the > > > > remaining: > > > > > > > > > > hive.metastore.uris > > > > > hive.metastore.warehouse.dir > > > > > fs.default.name > > > > > > > > > > George > > > > > > > > > > -- > > > > > -- > > > > > "Not everything that can be counted counts, and not everything that > > > > counts > > > > > can be counted." Albert Einstein > > > > > > > > > > > > > > > > > > > > > -- > > > -- > > > "Not everything that can be counted counts, and not everything that > > counts > > > can be counted." Albert Einstein > > > > > > > > > -- > -- > "Not everything that can be counted counts, and not everything that counts > can be counted." Albert Einstein >
