Example: Enabling the PostgreSQL Software to Run in the Cluster

Create the directories for the databases, WAL archives, configurations, utilities, and the log file.

Note - Perform the following steps in your target environment either in the global or in the local zone unless it is specified otherwise.

phys-schost-1$ mkdir /postgres/data

phys-schost-1$ mkdir /postgres/logs

phys-schost-1$ mkdir /postgres/83_walarchives

phys-schost-2$ mkdir /postgres/data

phys-schost-2$ mkdir /postgres/utilities

phys-schost-2$ mkdir /postgres/log

phys-schost-2$ mkdir /postgres/83_walarchives

Change to the PostgreSQL root directory and initialize the data cluster.

phys-schost-1$ cd /postgres/postgresql-8.3.1

phys-schost-1$ ./bin/initdb -D /postgres/data

phys-schost-2$ cd /postgres/postgresql-8.3.1

phys-schost-2$ ./bin/initdb -D /postgres/data

Start the database.

phys-schost-1$ ./bin/postmaster -D /postgres/postgresql-8.3.1

Prepare the Oracle Solaris Cluster specific test database.
Note - If you are in a local zone, ensure that you have access to a copy of your configuration file.
```
phys-schost-1$ ksh /opt/SUNWscPostgreSQL/util/pgs_db_prep -f 
\/postgres/pgs_config_pri -e
```

Stop the postmaster.

phys-schost-1$ ./bin/pg_ctl -D /postgres/data stop

Copy the PGDATA directory to the standby.
```
phys-schost-1$ cd /postgres
```
```
phys-schost-1$ /usr/local/bin/rysync -arv ./data phys-schost-2:/postgres
```
Note - If your target environment is in a zone, use zone 2 instead of phys-schost-2.

Protect the PostgreSQL configuration files.

Note - The PostgreSQL configuration files are overwritten during resilvering. You need to move the configuration files to prevent them from being overwritten.

phys-schost-1$ cd /postgres

phys-schost-1$ mkdir config

phys-schost-1$ cd data

phys-schost-1$ mv postgresql.conf ../config

phys-schost-1$ ln -s ../config/postgresql.conf ./postgresql.conf

phys-schost-1$ touch ../config/recovery.conf

phys-schost-1$ ln -s ../config/recovery.conf ./recovery.done

phys-schost-2$ cd /postgres

phys-schost-2$ mkdir config

phys-schost-2$ cd data

phys-schost-2$ mv postgresql.conf ../config

phys-schost-2$ ln -s ../config/postgresql.conf ./postgresql.conf

phys-schost-2$ touch ../config/recovery.conf

phys-schost-2$ ln -s ../config/recovery.conf ./recovery.conf

Provide the contents for the PostgreSQL recovery file.

phys-schost-1$ echo restore_command = 'cp /pgs/83_walarchives/%f %p' \
> /postgresql/data/recovery.done

phys-schost-2$ echo restore_command = '/postgres/postgres-8.3.1/bin \
/pg_standby -k 10 -t /postgres/data/failover /postgres/83_walarchives %f %p' \
/postgresql/data/recovery.conf

Configure the archive command in the postgresql.conf file on phys-schost-1 or zone1 by providing the following content.
```
archive_command = '/usr/local/bin/rsync -arv %p \
phys-schost-2:/postgres/83_walarchives/%f </dev/null'
```
Note - If you are in a zone environment , replace phys-schost-2 by zone 2 in the example.
Configure the archive command in the postgresql.conf file on phys-schost-2 or zone2 by providing the following content.
```
archive_command = '/usr/local/bin/rsync -arv %p \
phys-schost-1:/postgres/83_walarchives/%f </dev/null'
```
Note - If you are in a zone environment, replace phys-schost-1 by zone 1 in the example.
Exit from the postgres user ID.
```
phys-schost-1# exit
```

Encrypt the password of the monitoring user on all nodes, to run the PostgreSQL resources.

phys-schost-1# ksh /opt/SUNWscPostgreSQL/util/pgs_register -f \
 /postgres/pgs_config_pri -e

phys-schost-2# ksh /opt/SUNWscPostgreSQL/util/pgs_register -f \
/postgres/pgs_config_sta -e

Run the pgs_register script in the global zone to register the resources.

phys-schost-1# ksh /opt/SUNWscPostgreSQL/util/pgs_register -f /postgres/pgs_config_pri

phys-schost-2# ksh /opt/SUNWscPostgreSQL/util/pgs_register -f /postgres/pgs_config_sta

phys-schost-1# ksh /opt/SUNWscPostgreSQL/rolechg/util/rolechg-register -f \
/postgres/rolechg_config

Add the following line to the postgresql.conf file in the PGDATA directory on both nodes and zones.
```
listen_addresses = 'localhost, ha-host1'
```
Add the following line to the pg_hba.conf file in the PGDATA directory on both nodes and zones.
```
host all all 0.0.0.0/0 password
```

Enable the resources in the global zone.

phys-schost-1# clresource enable STA-RS

phys-schost-1# clresource enable PRIM-RS

phys-schost-1# clresource enable ROLECHG-RS

Copy the resilver scripts by performing the following steps in your target environment, either in the global or local zone.

phys-schost-2# cp /opt//SUNWscPostgreSQL/rolechg/util/resilver-step1 \
/postgres/utilities

phys-schost-2# cp /opt//SUNWscPostgreSQL/rolechg/util/resilver-step2 \
/postgres/utilities

phys-schost-2# chown -R postgres:postgres /postgres/utilities

Modify the following variables in your copy of the resilver-step1 script.

##### Customize the following variables##########


SOURCE_DATA=/postgres/data
TARGET_DATA=/postgres/data
TARGET=phys-schost-1
PGS_BASE=/postgres/postgresql-8.3.1
PRI_GRP=POSTGRES-PRI-RG
STDBY_GRP=POSTGRES-STA-RG
STDBY_RS=STA-RS
PGPORT=5432
ROLECHG_GRP=ROLECHG-RG
RESYNC=/usr/local/bin/rsync -rav
SSH_PASSPHRASE=false



##### End of customizations ##########

Modify the following variables in your copy of the resilver-step2 script.

##### Customize the following variables##########

SOURCE=phys-schost-2
SOURCE_DATA=/postgres/data
TARGET_DATA=/postgres/data
TARGET=phys-schost-1
PGS_BASE=/postgres/postgresql-8.3.1
PRI_GRP=POSTGRES-PRI-RG
STDBY_GRP=POSTGRES-STA-RG
STDBY_RS=STA-RS
PGPORT=5432
ROLECHG_GRP=ROLECHG-RG
PRI_NODE=phys-schost-1
RESYNC=/usr/local/bin/rsync -rav
SSH_PASSPHRASE=false



##### End of customizations ##########

Skip Navigation Links
Exit Print View
	Oracle Solaris Cluster Data Service for PostgreSQL Guide Oracle Solaris Cluster 3.3 3/13