thanks for the help guys. note the following:
* beeline -u jdbc:hive2:// works for engine=tez (local client) * beeline -u jdbc:hive2://dwrdevnn1.sv2.trulia.com:10001/default does not. (HS2 client) * both work for engine=mr so it does appear tez related. * nothing gets submitted to yarn. * i used Gopal's fragment for tez-site.xml * i've restarted the hiveserve2 process with VERBOSE logging * i've bounced the yarn RM. still getting the hang action. I'll post over to user@hive the log snippet as suggested. thanks again! Cheers, Stephen. On Wed, Feb 17, 2016 at 2:51 PM, Bikas Saha <[email protected]> wrote: > Gopal, > > Which of these configs can be updated in tez defaults so that we have them > out of the box? > > Bikas > > -----Original Message----- > From: Gopal Vijayaraghavan [mailto:[email protected]] On Behalf Of > Gopal > Vijayaraghavan > Sent: Wednesday, February 17, 2016 1:13 PM > To: [email protected] > Cc: Stephen Sprague <[email protected]> > Subject: Re: tez and beeline > > > set mapreduce.framework.name > > That can cause queries to fail entirely. > > Similarly with hive.exec.parallel=true. > > > things just hang at this point. > > You can pull up the HiveServer logs for that (post over to user@hive > lists & > I will follow there). > > > Any recommendations that i need to make to get things to work via HS2? > > You need to setup a bunch of configs in hive-site.xml & needs a valid > tez-site.xml for Tez (tez libs uploaded to HDFS etc) in general. > > For instance, I maintain recent versions of configs in my github > > https://github.com/t3rmin4t0r/tez-autobuild/blob/llap/tez-site.xml.frag > > + > https://github.com/t3rmin4t0r/tez-autobuild/blob/llap/hive-site.xml.frag > > > But after those configs, beeline + hs2 usually works fine. > > Cheers, > Gopal > PS: assuming you've loaded data via "INSERT", turning on > hive.compute.query.using.stats=true should skip a Tez job for a count(*) > > > >
