If you're interested in something similar to the Hive Server, but more threadsafe, check out the "Beeswax" server: http://github.com/cloudera/hue/tree/master/apps/beeswax . We open-sourced Hue (formerly Cloudera Desktop) yesterday, and one of the available (open source, Apache licensed) applications is Beeswax, which provides a UI that can run multiple Hive queries simultaneously. A screenshot is available at http://twitpic.com/21ae2b .
To get an idea of what's going on, take a look at the Thrift interface file: http://github.com/cloudera/hue/blob/master/apps/beeswax/thrift/beeswax.thrift . We did some work in the server implementation to identify the appropriate Hive thread locals and to do the right things with them. Cheers, -- Philip On Fri, Jun 25, 2010 at 9:28 AM, Omer, Farah <[email protected]> wrote: > I see. Thanks all, for the response. > > - Farah > > -----Original Message----- > From: Edward Capriolo [mailto:[email protected]] > Sent: Friday, June 25, 2010 10:39 AM > To: [email protected] > Subject: Re: Hive and ODBC driver- single threaded? > > On Fri, Jun 25, 2010 at 10:33 AM, Amogh Vasekar <[email protected]> wrote: >> Hi, >> Incidentally was looking into a similar thing. The Hive server is not >> threadsafe, >> >> https://issues.apache.org/jira/browse/HIVE-187?focusedCommentId=12738494&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#action_12738494 >> >> for more. >> >> Amogh >> >> >> On 6/25/10 7:16 PM, "Omer, Farah" <[email protected]> wrote: >> >> Hi all, >> >> I am having a question about Hive and the Hive ODBC driver (with latest >> patch HIVE-1101 applied). I was wondering if anybody can tell me if the ODBC >> driver is single-threaded or if Hive itself is single-threaded? The ODBC >> driver probably is, but I am wondering about Hive itself. I.e. can we submit >> and run two jobs concurrently? >> >> Any response much appreciated. Thanks. >> >> Farah Omer >> Senior DB Engineer, MicroStrategy, Inc. >> >> T: 703 2702230 >> E: [email protected] <mailto:[email protected]> >> http://www.microstrategy.com <http://www.microstrategy.com> >> >> >> >> >> > > Hadoop & Hive have some static thread local variables that propagate > those issues do the Web Interface Hive Server, JDBC. The simple (yet > annoying) solution is to spawn off multiple HiveServers. > > Edward >
