Hi sky,
Do you want to use customized hadoop cluster but not the mini cluster? For example, testing the latest version of Impala upon your existing Hive cluster. If so, you can modify the configuration files in ./fe/src/test/resources. They're just symbolic links. You can link them to your hadoop configuration files. Then the impala cluster will use your hadoop cluster. At 2017-07-18 18:14:32, "sky" <x_h...@163.com> wrote: > After I changed the impala-config.sh file, the HADOOP_CONF_DIR variables > did not take effect. This is still the case after running this > file(./bin/create-test-configuration.sh). >It seems that this variable does not take effect, because I only put hadoop >configuration file into the ./bin/start-impala -cluster.sh file of the same >directory will recognize the configuration. So I want to know how to do so >that any startup is not necessarily in the hadoop configure path. > > > > > > > >At 2017-07-18 02:33:11, "Tim Armstrong" <tarmstr...@cloudera.com> wrote: >>I'm not sure that I fully understand the question. >> >>There isn't a way to override HADOOP_CONF_DIR mostly - most scripts source >>impala-config.sh. >> >>On Sun, Jul 16, 2017 at 8:31 PM, sky <x_h...@163.com> wrote: >> >>> Hi Tim, >>> I found it from ./bin/create-test-configuration.sh that generating >>> ./fe/src/test/resources configurations, and HADOOP_CONFIG_DIR variable also >>> points to this directory. But I change this variable is not take effect. Is >>> this a hard code?
