Quick Start Guide for Apache Hawq［作者：盛想］

Apache HAWQ became an Apache open source project several weeks ago. Today I will introduce how we install and start Apache HAWQ.

Basic Requirements

Basic requirements has two parts, basic environments and some dependencies.

The first thing is to get a redhat 6.5 server and install the development tools during the setup of the system.

There are some dependencies you must install before the compiling. You should compile the libraries’ and install the libraries. Most libraries should compile and install to your system using the basic several steps :

       cd $LIBRARY_CODE_HOME

      ./configure

       make

Become root user and run :

      make install

This way would install the libraries` headers to /usr/local/include and libraries file to /usr/local/lib or /usr/local/lib64. If you want to install the libraries in another place, you can use the --prefix=/the/path option when you run configure. For example:

      ./configure --prefix=/the/path/you/want

But don`t forget to add the /the/path/you/want to your LD_LIBRARY_PATH in case of HAWQ cannot find the libraries while compiling.

We have a libraries list that you should install before compile HAWQ. The list contains the library name , version and download url.

Name	Version	Download Url
gperf	3.0.4	http://ftp.gnu.org/gnu/gperf/gperf-3.0.4.tar.gz
snappy	1.1.3	https://github.com/google/snappy/releases/download/1.1.3/snappy-1.1.3.tar.gz
bzip2	1.0.6	http://www.bzip.org/1.0.6/bzip2-1.0.6.tar.gz
python	2.6.2	https://www.python.org/ftp/python/2.6.2/Python-2.6.2.tgz
libevent	1.4.6	https://github.com/downloads/libevent/libevent/libevent-1.4.6-stable.tar.gz
json-c	0.9	http://oss.metaparadigm.com/json-c/json-c-0.9.tar.gz
krb5	1.11.3	http://web.mit.edu/kerberos/dist/krb5/1.11/krb5-1.11.3-signed.tar
libuuid	2.26.2	http://mirrors.163.com/centos/6.7/os/x86_64/Packages/libuuid-devel-2.17.2-12.18.el6.x86_64.rpm
libgsasl	1.8.0	ftp://ftp.gnu.org/gnu/gsasl/libgsasl-1.8.0.tar.gz
libxml2	2.7.8	http://xmlsoft.org/sources/libxml2-2.7.8.tar.gz
zlib	1.2.3	http://zlib.net/fossils/zlib-1.2.3.tar.gz
readline	6	ftp://ftp.gnu.org/gnu/readline/readline-6.0.tar.gz
openssl	0.9.8	ftp://ftp.openssl.org/source/openssl-0.9.8ze.tar.gz
protobuf	2.5.0	https://github.com/google/protobuf/tree/v2.5.0
boost	1.56	http://sourceforge.net/projects/boost/files/boost/1.56.0/boost_1_56_0.tar.bz2
bison	2.5	http://ftp.gnu.org/gnu/bison/bison-2.5.tar.gz
thrift	0.9.1-1	http://archive.apache.org/dist/thrift/0.9.1/thrift-0.9.1.tar.gz
apr	1.2.12	http://archive.apache.org/dist/apr/apr-1.2.12.tar.gz
libyaml	0.1.1	http://pyyaml.org/download/libyaml/yaml-0.1.1.tar.gz
curl	7.44.0	http://www.curl.haxx.se/download/curl-7.44.0.tar.gz
flex	2.5.35	http://sourceforge.net/projects/flex/files/flex-2.5.35.tar.gz/download
libhdfs3		git clone https://github.com/PivotalRD/libhdfs3.git  Install libhdfs3 follow the steps specified in README.md

Get Source Code

The Apache HAWQ source code can be obtained from Github. Click "Read More" in the bottom of this article to view the source code of HAWQ on Github. You could get the source code using git clone :

      git clone https://github.com/apache/incubator-hawq.git

Install Procedure

The source code includes libyarn under $HAWQ_CODE_HOME/depends directory. And you need to compile and install it by hand. Change directory to libyarn folder and read README

      cd $HAWQ_CODE_HOME/depends/libyarn/

To build libyarn, some libraries listed in the README are needed. You should install the libraries first and compile libyarn follow the steps specified in the README.

The first step of the building is to check whether you have installed all the dependencies. This is done by running the configure script.This script will run a number of tests to determine values for various system dependent variables and finally create the makefile. For a default installation simply run:

      ./configure

You could use --prefix=/hawq/install/path to change the HAWQ install path. Otherwise HAWQ would be install to /usr/local/ .

After finished configure, you can start to build HAWQ.To start the build, run:

      make

To build concurrently, run make with -j option.For example :

      make -j8

The build will take a few minutes depending on your hardware. The last line displayed should be:

      All of HAWQ successfully made. Ready to install.

Become root user and run :

      make install

You have installed HAWQ to your system !

Start Apache HAWQ

Since HAWQ bases on HDFS, you should ensure HDFS work before we init HAWQ. You could refer details from : http://hawq.docs.pivotal.io/docs-hawq/topics/InstallingHAWQ.html#preparingtoinstallhawq.

Now you can init and start HAWQ.

      cd $APACHE_HAWQ_HOME

      source greenplum_path.sh

      hawq init cluster

Once you have inited successful, you can start HAWQ using:

      hawq start cluster

To enter the database:

      psql -d postgres

You can run the create database query to test that HAWQ is running:

[gpadmin@localhost ~]$ psql -d postgres

psql (8.2.15)

Type "help" for help.

postgres=# create database tpch;

CREATE DATABASE

postgres=# \c tpch

You are now connected to database "tpch" as user "gpadmin".

tpch=# create table t(i int);

CREATE TABLE

tpch=# \timing

Timing is on.

tpch=# insert into t select generate_series(1,100);

INSERT 0 100

Time: 166.311 ms

tpch=# select count(*) from t;

count

-------

100

(1 row)

Time: 127.698 ms

Basic Operations

When you want to stop hawq cluster, run:

      hawq stop cluster

Or to restart the cluster:

      hawq restart cluster

The usage of command hawq is

      hawq <command> [<object>] [options]

The most commonly used hawq "commands" are:

      start              Start hawq service

      stop              Stop hawq service

      init                 Init hawq service

      restart           Restart hawq service

The object are:

      cluster            Operate hawq cluster.

      master            Operate hawq master.

      segment         Operate local segment node.

      standby          Operate hawq standby.

      allsegments    Operate all segments.

If you want to know all the hawq command , just run hawq or hawq --help. You can find all the usage of hawq command.