You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@sqoop.apache.org by "Masatake Iwasaki (Updated) (JIRA)" <ji...@apache.org> on 2012/03/27 13:44:28 UTC
[jira] [Updated] (SQOOP-390) PostgreSQL connector for direct export
with pg_bulkload
[ https://issues.apache.org/jira/browse/SQOOP-390?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Masatake Iwasaki updated SQOOP-390:
-----------------------------------
Description:
h1. Features
* Fast data export into PostgreSQL database with [pg_bulkload|http://pgbulkload.projects.postgresql.org/index.html].
* User can get benefit of functionality of pg_bulkload such as
** fast export bypassing shared bufferes and WAL,
** removing invalid data which cause parse error,
** ETL feature with filter functions.
h1. Implementation
h2. PGBulkloadExportMapper
At first, each map tasks create their own staging table with names based on task attempt id because pg_bulkload holds table level lock.
Arguments of pg_bulkload command can be passed via configuration object.
The Mapper export data by invoking pg_bulkload command as external process.
Commnad execution is done in the same way as DirectPostgresqlManager.
h2. PGBulkloadExportReducer
Reducer migrates data from staging tables into destination table.
Reducer gets the names of staging tables as map output values.
In order to do migration in a transaction, the number of reducers must be 1. (It is set by ConnectionManager internally).
Migration is done in same way as Sqoop defalt connectors using "INSERT INTO dst ( SELECT * FROM src )".
In the cleanup method, staging tables are dropped.
If exception is raised in the reducer, garbage staging tables are left.
User can delete them by executing this connector with --clear-staging-table option.
h1. Requirements
* pg_bulkload must be installed on DB server and all slave nodes.
** RPM for RedHat or CentOS is available in [download page|http://pgfoundry.org/frs/?group_id=1000261].
* [PostgreSQL JDBC| http://jdbc.postgresql.org/download.html] is also required on client node (same as PostgresqlManager).
* Superuser role of PostgreSQL database is required for execution of pg_bulkload.
h1. Usage
Currently there is no Factory class.
Specify connection manager class name with --connection-manager option to use.
{noformat}
$ sqoop export --connect jdbc:postgresql://localhost:5432/test \
--connection-manager com.cloudera.sqoop.manager.PGBulkloadManager \
--table test --username postgres --export-dir=/test -m 1
{noformat}
You can also specify pg_bulkload configuration with Hadoop configuration properties.
{noformat}
$ sqoop export \
-Dpgbulkload.bin="/usr/local/bin/pg_bulkload" \
-Dpgbulkload.input.field.delim=$'\t' \
-Dpgbulkload.check.constraints="YES" \
-Dpgbulkload.parse.errors="INFINITE" \
-Dpgbulkload.duplicate.errors="INFINITE" \
--connect jdbc:postgresql://localhost:5432/test \
--connection-manager com.cloudera.sqoop.manager.PGBulkloadManager \
--table test --username postgres --export-dir=/test -m 1
{noformat}
h1. Test
There is test class named PGBulkloadManagerTest extending TestExport.
This test expects that
* The major version of PostgreSQL is 9.0,
* The version of pg_bulkload is 3.1.1,
* PostgreSQL server is running on localhost:5432 of testing node,
* there is database named sqooptest,
* super user role named sqooptest exists,
* no password is set for the role, or .pgpass file is created.
Tests can be invoked as below.
{noformat}
# ant -Dtestcase=PGBulkloadManagerTest test
{noformat}
h2. Test on single-node cluster built on CentOS 6.0
Hadoop is expected to be running in pseudo distributed mode.
{noformat}
# rpm -ivh jdk-6u29-linux-amd64.rpm
# yum install ant-junit
# yum install hadoop-0.20-conf-pseudo
# sudo -u hdfs hadoop namenode -format
# /etc/init.d/hadoop-0.20-namenode start
# /etc/init.d/hadoop-0.20-datanode start
# /etc/init.d/hadoop-0.20-jobtracker start
# /etc/init.d/hadoop-0.20-tasktracker start
{noformat}
Setup for PostgreSQL:
{noformat}
# wget http://ftp.postgresql.org/pub/source/v9.0.7/postgresql-9.0.7.tar.bz2
# tar jxf postgresql-9.0.7.tar.bz2
# cd postgresql-9.0.7
# ./configure --prefix=/usr/local
# make
# make install
# useradd postgres
# mkdir /var/pgdata
# chown postgres:postgres /var/pgdata
# chmod 700 /var/pgdata
# sudo -u postgres /usr/local/bin/initdb -D /var/pgdata -E utf-8
# sudo -u postgres /usr/local/bin/pg_ctl start -D /var/pgdata -l /var/pgdata/postgresql.log
# createuser -U postgres -s sqooptest
# createdb -U sqooptest sqooptest
{noformat}
Setup for pg_bulkload:
{noformat}
# cd ..
# wget http://pgfoundry.org/frs/download.php/3176/pg_bulkload-3.1.1.tar.gz
# tar zxf pg_bulkload-3.1.1.tar.gz
# mv pg_bulkload-3.1.1 postgresql-9.0.7/contrib/
# cd postgresql-9.0.7/contrib/pg_bulkload-3.1.1
# make
# make install
# psql -U sqooptest -f /usr/local/share/postgresql/contrib/pg_bulkload.sql sqooptest
# ln -s /usr/local/bin/pg_bulkload /usr/bin/
{noformat}
Setup for PostgreSQL JDBC Driver:
{noformat}
# wget http://jdbc.postgresql.org/download/postgresql-9.0-802.jdbc4.jar
# cp postgresql-9.0-802.jdbc4.jar /usr/lib/hadoop/lib/
{noformat}
run test.
{noformat}
# cd /usr/local/src/sqoop-trunk
# ant -Dtestcase=PGBulkloadManagerTest test
{noformat}
was:
h1. Features
* Fast data export with pg_bulkload.
** http://pgbulkload.projects.postgresql.org/index.html
* User can get benefit of functionality of pg_bulkload such as
** bypassing shared bufferes and WAL,
** ignoring data causing parse error,
** ETL feature with filter functions.
h1. Implementation
* Each map tasks create their own staging table with names based on task attempt id. Then export data with pg_bulkload invoked as outer process.
* Staging tables are erased on
** Successful completion of export,
** or Exception in map task.
* Reduce task migrate data from staging tables into destination table
** Number of reduce tasks is internally set to 1 .
h1. Requirements
* pg_bulkload must be installed on DB server and all slave nodes.
* Also PostgreSQL JDBC is required on client. (Same as PostgresqlManager)
* Superuser role of PostgreSQL database is required for pg_bulkload.
h1. Usage
Currently there is no Factory class. Specify connection manager class name with --connection-manager option to use.
{noformat}
sqoop export --connect jdbc:postgresql://localhost:5432/test \
--connection-manager com.cloudera.sqoop.manager.PGBulkloadManager \
--table test --username postgres --export-dir=/test -m 1
{noformat}
You can also specify pg_bulkload configuration with Hadoop Configuration properties.
{noformat}
-Dpgbulkload.bin="/usr/bin/pg_bulkload"
-Dpgbulkload.input.field.delim=$'\t'
-Dpgbulkload.check.constraints="YES"
-Dpgbulkload.parse.errors="INFINITE"
-Dpgbulkload.duplicate.errors="INFINITE"
{noformat}
h1. Test
There is test class named PGBulkloadManagerTest extending TestExport.
{noformat}
ant -Dtestcase=PGBulkloadManagerTest test
{noformat}
This test requires
* PostgreSQL running on localhost:5432,
* database named sqooptest,
* super user role named sqooptest with no password, or .pgpass created
> PostgreSQL connector for direct export with pg_bulkload
> -------------------------------------------------------
>
> Key: SQOOP-390
> URL: https://issues.apache.org/jira/browse/SQOOP-390
> Project: Sqoop
> Issue Type: New Feature
> Reporter: Masatake Iwasaki
> Attachments: SQOOP-390-1.patch
>
>
> h1. Features
> * Fast data export into PostgreSQL database with [pg_bulkload|http://pgbulkload.projects.postgresql.org/index.html].
> * User can get benefit of functionality of pg_bulkload such as
> ** fast export bypassing shared bufferes and WAL,
> ** removing invalid data which cause parse error,
> ** ETL feature with filter functions.
> h1. Implementation
> h2. PGBulkloadExportMapper
> At first, each map tasks create their own staging table with names based on task attempt id because pg_bulkload holds table level lock.
> Arguments of pg_bulkload command can be passed via configuration object.
> The Mapper export data by invoking pg_bulkload command as external process.
> Commnad execution is done in the same way as DirectPostgresqlManager.
> h2. PGBulkloadExportReducer
> Reducer migrates data from staging tables into destination table.
> Reducer gets the names of staging tables as map output values.
> In order to do migration in a transaction, the number of reducers must be 1. (It is set by ConnectionManager internally).
> Migration is done in same way as Sqoop defalt connectors using "INSERT INTO dst ( SELECT * FROM src )".
> In the cleanup method, staging tables are dropped.
> If exception is raised in the reducer, garbage staging tables are left.
> User can delete them by executing this connector with --clear-staging-table option.
> h1. Requirements
> * pg_bulkload must be installed on DB server and all slave nodes.
> ** RPM for RedHat or CentOS is available in [download page|http://pgfoundry.org/frs/?group_id=1000261].
> * [PostgreSQL JDBC| http://jdbc.postgresql.org/download.html] is also required on client node (same as PostgresqlManager).
> * Superuser role of PostgreSQL database is required for execution of pg_bulkload.
> h1. Usage
> Currently there is no Factory class.
> Specify connection manager class name with --connection-manager option to use.
>
> {noformat}
> $ sqoop export --connect jdbc:postgresql://localhost:5432/test \
> --connection-manager com.cloudera.sqoop.manager.PGBulkloadManager \
> --table test --username postgres --export-dir=/test -m 1
> {noformat}
> You can also specify pg_bulkload configuration with Hadoop configuration properties.
> {noformat}
> $ sqoop export \
> -Dpgbulkload.bin="/usr/local/bin/pg_bulkload" \
> -Dpgbulkload.input.field.delim=$'\t' \
> -Dpgbulkload.check.constraints="YES" \
> -Dpgbulkload.parse.errors="INFINITE" \
> -Dpgbulkload.duplicate.errors="INFINITE" \
> --connect jdbc:postgresql://localhost:5432/test \
> --connection-manager com.cloudera.sqoop.manager.PGBulkloadManager \
> --table test --username postgres --export-dir=/test -m 1
> {noformat}
> h1. Test
> There is test class named PGBulkloadManagerTest extending TestExport.
> This test expects that
> * The major version of PostgreSQL is 9.0,
> * The version of pg_bulkload is 3.1.1,
> * PostgreSQL server is running on localhost:5432 of testing node,
> * there is database named sqooptest,
> * super user role named sqooptest exists,
> * no password is set for the role, or .pgpass file is created.
> Tests can be invoked as below.
> {noformat}
> # ant -Dtestcase=PGBulkloadManagerTest test
> {noformat}
> h2. Test on single-node cluster built on CentOS 6.0
> Hadoop is expected to be running in pseudo distributed mode.
> {noformat}
> # rpm -ivh jdk-6u29-linux-amd64.rpm
> # yum install ant-junit
> # yum install hadoop-0.20-conf-pseudo
> # sudo -u hdfs hadoop namenode -format
> # /etc/init.d/hadoop-0.20-namenode start
> # /etc/init.d/hadoop-0.20-datanode start
> # /etc/init.d/hadoop-0.20-jobtracker start
> # /etc/init.d/hadoop-0.20-tasktracker start
> {noformat}
> Setup for PostgreSQL:
> {noformat}
> # wget http://ftp.postgresql.org/pub/source/v9.0.7/postgresql-9.0.7.tar.bz2
> # tar jxf postgresql-9.0.7.tar.bz2
> # cd postgresql-9.0.7
> # ./configure --prefix=/usr/local
> # make
> # make install
> # useradd postgres
> # mkdir /var/pgdata
> # chown postgres:postgres /var/pgdata
> # chmod 700 /var/pgdata
> # sudo -u postgres /usr/local/bin/initdb -D /var/pgdata -E utf-8
> # sudo -u postgres /usr/local/bin/pg_ctl start -D /var/pgdata -l /var/pgdata/postgresql.log
> # createuser -U postgres -s sqooptest
> # createdb -U sqooptest sqooptest
> {noformat}
> Setup for pg_bulkload:
> {noformat}
> # cd ..
> # wget http://pgfoundry.org/frs/download.php/3176/pg_bulkload-3.1.1.tar.gz
> # tar zxf pg_bulkload-3.1.1.tar.gz
> # mv pg_bulkload-3.1.1 postgresql-9.0.7/contrib/
> # cd postgresql-9.0.7/contrib/pg_bulkload-3.1.1
> # make
> # make install
> # psql -U sqooptest -f /usr/local/share/postgresql/contrib/pg_bulkload.sql sqooptest
> # ln -s /usr/local/bin/pg_bulkload /usr/bin/
> {noformat}
> Setup for PostgreSQL JDBC Driver:
> {noformat}
> # wget http://jdbc.postgresql.org/download/postgresql-9.0-802.jdbc4.jar
> # cp postgresql-9.0-802.jdbc4.jar /usr/lib/hadoop/lib/
> {noformat}
> run test.
> {noformat}
> # cd /usr/local/src/sqoop-trunk
> # ant -Dtestcase=PGBulkloadManagerTest test
> {noformat}
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira