it turns out to be a environment issue. i was running hive server using another user: sudo -u op but the system resets all environment variables for that user under sudo command. the fix i put in is to put environment variable names in env_keep in /etc/sudoers. after that hive server is running in remote mode now.
thx guys for the feedback, :) On Sat, Jun 11, 2011 at 5:07 AM, sangeetha s <sangee....@gmail.com> wrote: > I havn't tried setting hive with derby server yet. I am using mysql as > metastore. But not much difference I think. > > To setup Hive server in remote machine you should add the all the hadoop > and hive packages (folders) under ur working directory. Set the hadoop > master and slaves > > accordingly. Setup the environment variable for hadoop_home, Hive_home, > java_home in ur .bashrc file. Make sure that the hadoop cluster is up and > running before > > starting hive. In the remote node start the hive server using > $HIVE_HOME/bin/ hive --service hiveserver. > > If you have configured derby then it will communicate it accordingly. > > Jus give a try. > > > On Sat, Jun 11, 2011 at 9:46 AM, Shouguo Li <the1plum...@gmail.com> wrote: > >> hmm, nobody replied to my email earlier so going to send this out one last >> time, sorry for the spam guys. >> btw, i also exported these env variables before starting hive server: >> export HADOOP_CONF_DIR >> export HADOOP_HOME >> export HIVE_CONF_DIR >> export HIVE_HOME >> export PATH=$HIVE_HOME/bin:$PATH >> thx! >> >> >> >> On Fri, Jun 3, 2011 at 2:29 PM, Shouguo Li <the1plum...@gmail.com> wrote: >> >>> hey guys >>> i added hive package to my hadoop cluster. if i go into hive cli, i can >>> run hive in remote mode. but queries going through hive server runs in local >>> mode which is really slow... the only changes i did to hive is to set up >>> derby network server, following this link: >>> http://wiki.apache.org/hadoop/HiveDerbyServerMode >>> anybody knows how to fix this? >> >> >> > > > -- > > > > Regards, > Sangita S >
