This information describes how to add the Db2® Big
SQL service on a Cloudera Distribution Including Apache Hadoop (CDH) cluster.
Before you begin
Make sure that you follow the prerequisite steps that are listed in Requirements for installing Db2 Big SQL.
Recommended: Manually run the
Db2 Big SQL pre-installation check utility on the candidate
Db2 Big SQL Head node, and passing in all other candidate Db2 Big SQL nodes. Review and address any issues before you
install Db2 Big SQL.
To run this utility, you must be logged on as root, a non-root user with sudo privileges, or the
Db2® Big SQL service user. In this release, logging on as a non-default bigsql
user is not supported.
If HBase is installed on your cluster and you want Db2 Big
SQL to access HBase, you must install each worker on either an HBase Region Server node or an HBase
client
node.
When you install Db2 Big SQL, there are guidelines that you
can follow to optimize performance. For more information, see Cluster and system tuning.
Procedure
-
On the candidate Db2 Big SQL Head node, create the host
list file /tmp/bigsqlHostList that contains the fully qualified domain name
(FQDN) of each node in your cluster.
Put each host name on a single line, starting with the target installation node for the Db2 Big SQL Head. The remaining host names become Db2 Big SQL worker nodes.
- Edit the installation configuration parameters to match your cluster environment by
running the Db2 Big
SQL configuration utility.
- Type the following commands to view the configuration parameters and their default
values:
cd /usr/ibmpacks/IBM-Big_SQL/5.0.4.0/bigsql-cli/
./bigsql-config -display
- Modify installation parameters that are not accurate for your cluster.
./bigsql-config -set "key=value"
At a minimum, modify the following configuration parameters:
For CM_HOST, specify the fully qualified domain name (FQDN) of the host. For HDFS_USER, make sure
that the specified value is a valid HDFS superuser. If the cluster is kerberized, ensure that
kinit
was run in the HDFS superuser account. The Db2 Big SQL installer uses the HDFS superuser to create a directory on HDFS.
For
example,
./bigsql-config -set "CM_HOST=<your.cloudera.manager.server.com>"
- Using the same utility, update configuration for the Db2 Big SQL service.
You can enable or disable the following options:
- HBase (if HBase is installed)
- Impersonation
- Pluggable Authentication Module (PAM)
For more information about these options, run the following command:
bigsql-config -help
Note: You can also do this step after Db2 Big SQL is
installed.
- Install Db2 Big SQL by running the
bigsql-install utility.
cd /usr/ibmpacks/IBM-Big_SQL/5.0.4.0/bigsql-cli/
./bigsql-install
- When prompted, enter passwords for the bigsql and Cloudera Manager
admin users.
Note: If the cluster is kerberized, Db2 Big SQL is
automatically kerberized during the installation process. You can manage the Db2 Big SQL service keytabs in Cloudera Manager.
- If Sentry is enabled on the cluster, run the following commands from the Hive
server:
create role bigsql_role;
grant role bigsql_role to group <bigsql_group>;
grant all on server server1 to role bigsql_role with grant option;
<bigsql_group> is the group of the bigsql user. To
obtain <bigsql_group>, run the following command:
./bigsql-config -get SYSADM_GRP
- Validate your install with the Db2 Big SQL cluster administration utility.
cd /usr/ibmpacks/IBM-Big_SQL/5.0.4.0/bigsql-cli/
./bigsql-admin -smoke
Tip: To run data load tests, run the command
/usr/ibmpacks/IBM-Big_SQL/5.0.4.0/bigsql-cli/BIGSQL/package/scripts/bigsql-smoke.sh
-l
.
Results
If there is an error during the installation process, you can fix the error and run the
installation utility again. Review installation log files in the
/tmp/<bigsql_user>/logs directory (for example,
/tmp/bigsql/logs). You can run the utility as many times as needed. When you
rerun the utility, the utility resumes from the last successful step.