No, I specify the SQL server ip. More over I see in the hadoop logs that the jdo options that are printed by the HiveMain are correct, but they are ignored when the mapred job is executed. It looks like the mapred job picks up local hive-defaults.xml (on the task-tracker node) and uses it ignoring the actual properties passed to it. Any siggestions?
On Tue, Jul 10, 2012 at 8:07 PM, Harsh J <[email protected]> wrote: > Hey Boris, > > What do you specify as the hostname for your DB in the Hive > configuration? Do you specify 'localhost'? If so, it will not work in > a distributed cluster, as Oozie's actions are run as Jobs with tasks > executing on TT machines (remotely), and they can't all have > connections to localhost for a DB being valid logically. > > On Wed, Jul 11, 2012 at 8:22 AM, Boris Shulman <[email protected]> wrote: > > Hi my hive action fails with: > > > > Failing Oozie Launcher, Main class > > [org.apache.oozie.action.hadoop.HiveMain], exit code [10] > > > > > > javax.jdo.JDOFatalDataStoreException: java.net.ConnectException : > > Error connecting to server localhost on port 1527 with message > > Connection refused > > > > > > it looks like it can't pick-up the right hive-defaults.xml file. I have > > changed the hive metastore to a SQL db in every possible place > (hive-site, > > hive-defaults), I also provided hive-default via job-config element, but > it > > still using the default value that I can't find where it is coming from. > > > > Any suggestions where it can pick the value from (I am using oozie 3.2.0 > > and hive 0.7.1)? > > > > -- > Harsh J >
