Hi,
I'm installing Impala in a cluster environment. I'm using CDH 4 with latest
Impala package.
I have two-node cluster to start with.
Node-01 is hosting NN, JT, statestored, hive and mysql installation.
Node-02 is hosting Datanode, tasktracker, Impalad.
After starting statestored (on node 01), I tried to start impalad on *node
02* with the following command
GVLOG_v=1 nohup /usr/bin/impalad -state_store_host=<NODE-01> -nn=<NODE-01>
-nn_port=8020 -hostname=<NODE-02> -ipaddress=<<NODE-02 IP Address>
Looks like Impalad is creating a local derby log and metsstore_db directory
on the current launching dir.
It definitely says that it is not honoring my hive-site.xml.
My hive-site.xml is located on node-02 at /usr/lib/impala/conf/ :
ls -ltr /usr/lib/impala/conf/
total 12
-rw-r--r--. 1 root root 2530 Jan 23 19:39 hive-site.xml
-rw-r--r--. 1 root root 1347 Jan 23 19:52 core-site.xml
-rw-r--r--. 1 root root 1693 Jan 23 20:04 hdfs-site.xml
Few points:
1. I'm starting the process as myself, not as root or impala.
2. When i created user (hive) on mysql, I used this: CREATE USER
'hive'@'Node-01' IDENTIFIED BY 'MYPASSWD';
3. My node-02 hive-site.xml contains this:
<property>
<name>javax.jdo.option.ConnectionURL</name>
<value>jdbc:mysql://NODE-01/metastore</value>
</property>
<property>
<name>javax.jdo.option.ConnectionDriverName</name>
<value>com.mysql.jdbc.Driver</value>
</property>
3. My impalad log ar /tmp directory doesn't contain any error message.
Can someone please help me out to identify the reason of not connecting to
the configured metastore?
Regards,
Mohammad
--
