This document describes the steps required to build ApacheTrafodion software.
<TBD - should I change all occurances of Trafodion to Apache Trafodion?>
<TBD - need to fix links>
Supported Platforms
Red Hat 6.4 or Centos 6.4 versions are supported as development and production platforms.
Required Software
Install the following packages via yum install <package>
alsa-lib-devel | ant | ant-nodeps |
boost-devel | device-mapper-multipath | dhcp |
gcc-c++ | gd | glibc-devel.i686 |
graphviz-perl | gzip | java-1.7.0-openjdk-devel |
openldap-devel | libaio-devel | libibcm.i686 |
libibumad-devel | libibumad-devel.i686 | libiodbc |
libiodbc-devel | librdmacm-devel | librdmacm-devel.i686 |
log4cxx | log4cxx-devel | lua-devel |
lzo-minilzo | net-snmp-devel | net-snmp-perl |
openldap-clients | openldap-devel.i686 | openmotif |
openssl-devel.i686 | openssl-static | perl-Config-IniFiles |
perl-DBD-SQLite | perl-Config-Tiny | perl-Expect |
perl-IO-Tty | perl-Math-Calc-Units | perl-Params-Validate |
perl-Parse-RecDescent | perl-TermReadKey | perl-Time-HiRes |
protobuf-compiler | protobuf-devel | python-qpid |
python-qpid-qmf | qpid-cpp-client | qpid-cpp-client-devel |
qpid-cpp-client-ssl | qpid-cpp-server | qpid-cpp-server-ssl |
qpid-qmf | qpid-tools | readline-devel |
saslwrapper | sqlite-devel | flex |
libXext-devel | libX11-devel | libXau-devel |
unixODBC | unixODBC-devel | uuid-perl |
xinetd | xerces-c-devel | libcurl-devel |
libxml2-devel |
Example yum install script:
<TDB - should we include this yum command? It make things easier when setting up a new VM>
yum install alsa-lib-devel ant ant-nodeps boost-devel device-mapper-multipath dhcp gcc-c++ gd glibc-devel glibc-devel.i686 \ graphviz-perl java-1.7.0-openjdk-devel libaio-devel libibcm.i686 libibumad-devel libibumad-devel.i686 \
libiodbc libiodbc-devel librdmacm-devel librdmacm-devel.i686 log4cxx log4cxx-devel lua-devel lzo-minilzo net-snmp-devel \
net-snmp-perl openldap-clients openldap-devel.i686 openmotif openssl-devel openssl-devel.i686 openssl-static \
perl-Config-IniFiles perl-DBD-SQLite perl-Config-Tiny perl-Expect perl-IO-Tty perl-Math-Calc-Units \
perl-Params-Validate perl-Parse-RecDescent perl-TermReadKey perl-Time-HiRes protobuf-compiler \
protobuf-devel python-qpid python-qpid-qmf qpid-cpp-client qpid-cpp-client-devel qpid-cpp-client-ssl \
qpid-cpp-server qpid-cpp-server-ssl qpid-qmf qpid-tools readline-devel saslwrapper sqlite-devel \
unixODBC unixODBC-devel uuid-perl xinetd xerces-c-devel \
openldap-devel flex libcurl-devel libxml2-devel
Check the following before continuing
Java 1.7.x or greater must be installed. Ensure JAVA_HOME environment variable exists and set to your JDK installation.
- Install openssl version "OpenSSL 1.0.1e-fips" or greater. To check your version do: "openssl version".
- Install protocol buffers version "libprotoc 2.3.0" or greater. To check your version do: "protoc –version"
- Install Apache Maven version 3.0.5 or greater and add it to your PATH. to check your version do: "mvn --version"
- Download latest Apache Trafodion source from Apache Incubator https://github.com/apache/incubator-trafodion, un-tar the file into the <Trafodion download directory>
- Download, build and install additional development tools via Additional Build Tools.
Note: Trafodion source needs to be downloaded before installing additional build tools if you are planning on running the Trafodion supplied convenience script to download needed packages.
<TBD, Java and openssl are downloaded through Yum. It looks like protocol buffers are part of Centos/Redhat, Do I need to add instructions on how to download and setup Maven?>
Build Steps
Set up Trafodion configuration file
<TDB: change hardcoded version number 1.2.0 to be configurable>
Additional development tools are required before building Trafodion as described Additional Build Tools. A convenience script script exists that downloads, installs, and builds all these tools in a common directory. If this convenience script is not used or if any of these additional build tools are not found in the expected location, then the Trafodion configuration file needs to be updated. The Trafodion configuration file template is located in <download directory>/apache-trafodion-1.2.0-incubating/core/sqf/LocalSettingsTemplate.sh. To change values, copy this file to your root directory and change its name to .trafodion. Edit the .trafodion file and update according to the instructions. Be sure to change the location of your TOOLSDIR to your <tools installation directory>.
cp <traf download directory>/apache-trafodion-1.2.0-incubating/core/sqf/LocalSettingsTemplate.sh ~/.trafodion
For now, don't change locations for HADOOP_PREFIX, HBASE_HOME, and HIVE_HOME. There is a later step that describes this process.
Build Trafodion
You can use single node standalone Apache installation OR use a Trafodion supplied solution that encapsulates the building and testing in an isolated environment . If you plan to use an existing standalone Apache installation please follow instructions for "Build using standalone Apache installation". If you want to use your existing environment, please following instructions for "Build using local Hadoop" ".
Build using standalone Apache installation
If you wish to build Trafodion using an existing standalone Apache installation, follow instructions in this section. With this option, you can only build Trafodion. If you want to test Trafodion, you will need to create binary tar files and follow our installation instructions described here (Installation). Note: the Trafodion installer only supports Hortonworks and Cloudera distributions at this time.
Much of Trafodion is written in C++ and it uses libraries that are built as part of the Hadoop native libraries. Therefore, your Hadoop distribution must have these libraries available. To check this:
cd <hadoop installation directory>
ls lib/native
The list of lib/native code should include libhdfs.so
If you have not already done so, copy the Trafodion configuration file into your root directory.
cp <download directory>/ apache-trafodion-1.2.0-incubating/core/sqf/LocalSettingsTemplate.sh ~/.trafodion
Edit .trafodion and change locations for HADOOP_PREFIX, HBASE_HOME, and HIVE_HOME to point to your installed environment. Be sure to update the TOOLSDIR variable in this file to point to the location of the components installed via Additional Build Tools.
Note: Trafodion has only been tested with Hadoop 2.5.?, HBase 0.98.6, and Hive 0.13.1. If you have different versions, you might run into unexpected build errors.
Setup build environment
Start a new ssh session
Set the following environment variables, these are needed to create correct class paths when setting up the build environment.
export HADOOP_PREFIX=<location of Hadoop 2.5.?>
export HBASE_HOME=<location of hbase-0.98.6>
export HIVE_HOME=<location of apache-hive-0.13.1>
export THRIFT_INC_DIR=<location of thrift-0.9.0/include>
export THRIFT_LIB_DIR=<location of thrift-0.9.0/lib>
Make sure that TOOLSDIR is not set.
unset $TOOLSDIR
Do the build
cd <download directory>/apache-trafodion-1.2.0-incubating
source ./env.sh
Build using one of the following options:
make all (Build Trafodion, DCS, REST) OR
make package (Build Trafodion, DCS, REST, Client drivers) OR
make package-all (Build Trafodion, DCS, REST, Client drivers and Tests for all components
Verify build by executing the following statement, this script should return over 90 components
$MY_SQROOT/sql/script/sqvers
Build using local Hadoop
This section describes the steps to use the Trafodion installation script called 'install_local_hadoop' that encapsulates building the product and starting the Hadoop eco-system. This script uses a Cloudera distribution.
Make sure you have set up password less authentication. You should be able to "ssh localhost" without having to enter a password
Install Hadoop eco-system
Start a new ssh session and set environment
Set the TOOLSDIR environment variable to the location of the components installed via Additional Build Tools. Be sure to set the TOOLSDIR environment variable before invoking env.sh
export TOOLSDIR=<tools installation directory>
cd <download directory>/apache-trafodion-1.2.0-incubating
source ./env.sh
cd $MY_SQROOT/sql/scripts
Execute the script ‘install_local_hadoop’, see Note below regarding assigning ports
Verify that build completed by running the following command - should report: 6 java servers and 2 mysqld processes are running
$MY_SQROOT/sql/scripts/swstatus
Note:
The 'install_local_hadoop' script downloads Hadoop, HBase, Hive, and MySql jar files from the internet. To avoid this overhead, you can download the required files into a separate directory and set the environment variable MY_LOCAL_SW_DIST to point to this directory.
The following options are available with 'install_local_hadoop':
‘install_local_hadoop –p rand' — will start with any random port number between 9000 and 49000 OR
‘install_local_hadoop –p < specify a port # >' — will start with port number specified OR
‘install_local_hadoop' — will use default port numbers for all services
When this script completes, Hadoop, HBase, Hive, and MySql (used as Hive's metadata repository) have been installed and are started.
The 'install_local_hadoop' script also creates several helper scripts starting with "sw" in the $MY_SQROOT/sql/scripts directory including:
swhadoop swhbase swhdfs
swhive swjdbc swmysql
swmysqladmin swstartall swstarthadoop
swstarthbase swstartmysql swstatus
swstopall swstophadoop swstophbase
swstopmysql swuninstall_local_hadoop swurls.html
swyarn
To start/stop/check Hadoop environment using Trafodion supplied scripts, you can execute ‘swstartall’ , ‘swstopall’ and ‘swstatus’, and if you need to remove the installation, execute the 'swuninstall_local_hadoop'.
Do the build
Start a new ssh session and set environment
cd <download directory>/apache-trafodion-1.2.0-incubating
source ./env.sh
Build using one of the following options:
make all (Build Trafodion, DCS, REST) OR
make package (Build Trafodion, DCS, REST, Client drivers) OR
make package-all (Build Trafodion, DCS, REST, Client drivers and Tests for all components
Verify build by executing the following statement, this script should return over 90 components
$MY_SQROOT/sql/script/sqvers
Install Trafodion components
cd $MY_SQROOT/sql/scripts
Execute script, ‘install_traf_components’.
Test Trafodion build
If you built Trafodion using "Build using standard Apache installation", following the instructions in "Test using standard Apache installation". If you built Trafodion using the "Build using local Hadoop", follow the instructions in "Test using local Hadoop".
Test using standard Apache installation
<TBD - add instructions on how to .tar up the build. This section needs some work>
Note: All tar files are created in ‘distribution’ folder located at the very top level (incubator-trafodion) during the make step.
To install a custom Trafodion component you can set various environment variable to overwrite the default tar files found in distribution folder.
Environment variables supported by install_traf_component script are :
DCS_TAR — Specify the fully qualified path of DCS tar file
REST_TAR — Specify the fully qualified path of REST tar file
PHX_TAR — Specify the fully qualified path of Phoenix test tar file
CLIENT_TAR — Specify the fully qualified path of Trafodion client tar file
DCSTEST_TAR — Specify the fully qualified path of DCS tests tar file
Follow installation instructions found on the Apache Trafodion wiki site - (Installation ) but modify the step that request you to download tar files from the download directory to use your generated the tar files.
Test using local Hadoop
This section describes how to start up Trafodion and run operations. It assumes that you have previously run the 'install_local_hadoop' script.
The 'install_local_hadoop' script configures everything needed to run Trafodion. It also starts Hadoop, Hive, MySql, and HBase.
Before starting, please check your ulimit. Some of them may need to be increased. If incorrect values are used, Trafodion may not start.
<TDB: should I add instructions on how to do this?>
The following are the recommended values.
ulimit –a
- core file size (blocks, -c) 1000000
- data seg size (kbytes, -d) unlimited
- scheduling priority (-e) 0
- file size (blocks, -f) unlimited
- pending signals (-i) 515196
- max locked memory (kbytes, -l) 49595556
- max memory size (kbytes, -m) unlimited
- open files (-n) 32000
- pipe size (512 bytes, -p) 8
- POSIX message queues (bytes, -q) 819200
- real-time priority (-r) 0
- stack size (kbytes, -s) 10240
- cpu time (seconds, -t) unlimited
- max user processes (-u) 267263
- virtual memory (kbytes, -v) unlimited
- file locks (-x) unlimited
Start Trafodion
Start a new ssh session and set environment
cd <download directory>/apache-trafodion-1.2.0-incubating
source ./env.sh
cd $MY_SQROOT/sql/scripts
Perform a one-time step to set up run environment, <TBD, add more details on when this is required and rm ms.env>
Execute the script, ‘sqgen’
Execute the script, 'sqstart'
Note: In case of any issues and if there is a need to stop and restart a specific Trafodion component, you can use the component based start/stop scripts.
Component | Start script | Stop script |
For all of Trafodion | sqstart | sqstop |
For DCS (Database Connectivity Service) | dcsstart | dcsstop |
For REST server | reststart | reststop |
For LOB server | lobstart | lobstop |
For RMS server | rmsstart | rmsstop |
|
|
|
Checking the status of Trafodion and its components
There are several health check scripts that are available which will provide the status of Trafodion. They are :
sqcheck (For all of Trafodion)
dcscheck (For Database Connectivity Service)
rmscheck (For RMS Server)
Creating Trafodion metadata
Trafodion is up and running, you can now start up a SQL command line interface and initialize Trafodion
Start a new ssh session and set environment
cd <download directory>/apache-trafodion-1.2.0-incubating
source ./env.sh
cd $MY_SQROOT/sql/scripts
Execute the script 'sqlci' (direct to sql engine) or 'trafci' (uses DCS to connect to SQL engine)
Perform the following statements:
- initialize trafodion;
- get schemas;
- create table table1 (a int);
- invoke table1;
- insert into table1 values (1), (2), (3), (4);
- select * from table1;
- exit;
You are done and ready to go! <TBD: add reference to documentation>
Trafodion test packages
There are several helper scripts provided to run the tests for Trafodion components in your workspace. These scripts are generated based on the tar files that are made available during execution of install_traf_component script
swphoenix {t4 | t2 } — This script will run the Phoenix test using JDBC Type 4 driver or JDBC Type 2 driver
swjdbc — This script will run JDBC Type 4 tests
swpyodbc — This script will install Linux drivers run ODBC tests using Linux driver