forked from ISTI-ansible-roles/ansible-roles
.gitignore: Ignore .idea files.
library/roles/postgresql: pgpool-II, manage the configuration file as a template. Manage authentication and authorization.
This commit is contained in:
parent
670a98fa2a
commit
8a1edd5801
|
@ -55,12 +55,60 @@ psql_db_pwd: "We cannot save the password into the repository. Use another varia
|
|||
postgresql_pgpool_pkgs:
|
||||
- 'postgresql-{{ psql_version }}-pgpool2'
|
||||
|
||||
#psql_db_data:
|
||||
# Example of line needed to create a db, create the user that owns the db, manage the db accesses (used by iptables too). All the fields are mandatory.
|
||||
#- { name: '{{ psql_db_name }}', encoding: 'UTF8', user: '{{ psql_db_user }}', pwd: '{{ psql_db_pwd }}', roles: 'NOCREATEDB,NOSUPERUSER', extensions: [ 'postgis', 'pgpool_regclass', 'pgpool_recovery' ], allowed_hosts: [ 'xxx.xxx.xxx.xxx/32', 'yyy.yyy.yyy.yyy/32' ] }
|
||||
# Example of line needed to manage the db accesses (used by iptables too), without creating the db and the user. Useful, for example, to give someone access to the postgresql db
|
||||
#- { name: '{{ psql_db_name }}', user: '{{ psql_db_user }}', allowed_hosts: [ 'xxx.xxx.xxx.xxx/32', 'yyy.yyy.yyy.yyy/32' ] }
|
||||
|
||||
|
||||
# pgpool-II
|
||||
pgpool_pkgs:
|
||||
- pgpool2
|
||||
|
||||
pgpool_enabled: True
|
||||
pgpool_listen_addresses: 'localhost'
|
||||
pgpool_port: 5433
|
||||
pgpool_listen_backlog_multiplier: 2
|
||||
pgpool_pcp_user: admin
|
||||
# Define pcp_pwd in a vault file
|
||||
pgpool_pcp_listen_addresses: '*'
|
||||
pgpool_pcp_port: 9898
|
||||
#pgpool_backends:
|
||||
# - { id: 0, hostname: 'backend0', backend_port: '{{ psql_db_port }}', backend_weight: 1, backend_data_directory: '{{ psql_data_dir }}', backend_flag: 'ALLOW_TO_FAILOVER' }
|
||||
pgpool_enable_pool_hba: 'on'
|
||||
pgpool_pool_passwd: 'pool_passwd'
|
||||
pgpool_num_init_children: 32
|
||||
pgpool_max_pool: 4
|
||||
pgpool_child_life_time: 300
|
||||
pgpool_child_max_connections: 0
|
||||
pgpool_connection_life_time: 0
|
||||
pgpool_client_idle_limit: 0
|
||||
pgpool_log_destination: syslog
|
||||
pgpool_log_connections: 'on'
|
||||
pgpool_log_hostname: 'on'
|
||||
pgpool_log_statement: 'off'
|
||||
pgpool_log_per_node_statement: 'off'
|
||||
pgpool_debug_level: 0
|
||||
pgpool_replication_mode: 'on'
|
||||
pgpool_replicate_select: 'off'
|
||||
pgpool_insert_lock: 'on'
|
||||
pgpool_lobj_lock_table: ''
|
||||
pgpool_replication_stop_on_mismatch: 'off'
|
||||
pgpool_failover_if_affected_tuples_mismatch: 'off'
|
||||
pgpool_load_balance_mode: 'on'
|
||||
pgpool_ignore_leading_white_space: 'on'
|
||||
pgpool_white_function_list: ''
|
||||
pgpool_black_function_list: 'nextval,setval'
|
||||
pgpool_allow_sql_comments: 'on'
|
||||
pgpool_fail_over_on_backend_error: 'on'
|
||||
pgpool_relcache_expire: 3600
|
||||
pgpool_memory_cache_enabled: 'off'
|
||||
pgpool_memqcache_method: memcached
|
||||
pgpool_memqcache_memcached_host: localhost
|
||||
pgpool_memqcache_memcached_port: 11211
|
||||
pgpool_memqcache_expire: 0
|
||||
pgpool_memqcache_auto_cache_invalidation: 'on'
|
||||
|
||||
|
||||
# postgis
|
||||
|
@ -84,10 +132,3 @@ pg_backup_use_auth: "yes"
|
|||
pg_backup_pass_file: /root/.pgpass
|
||||
pg_backup_use_nagios: "yes"
|
||||
|
||||
#psql_db_data:
|
||||
# Example of line needed to create a db, create the user that owns the db, manage the db accesses (used by iptables too). All the fields are mandatory.
|
||||
#- { name: '{{ psql_db_name }}', encoding: 'UTF8', user: '{{ psql_db_user }}', pwd: '{{ psql_db_pwd }}', roles: 'NOCREATEDB,NOSUPERUSER', extensions: [ 'postgis', 'pgpool_regclass', 'pgpool_recovery' ], allowed_hosts: [ 'xxx.xxx.xxx.xxx/32', 'yyy.yyy.yyy.yyy/32' ] }
|
||||
# Example of line needed to manage the db accesses (used by iptables too), without creating the db and the user. Useful, for example, to give someone access to the postgresql db
|
||||
#- { name: '{{ psql_db_name }}', user: '{{ psql_db_user }}', allowed_hosts: [ 'xxx.xxx.xxx.xxx/32', 'yyy.yyy.yyy.yyy/32' ] }
|
||||
|
||||
|
||||
|
|
|
@ -6,3 +6,11 @@
|
|||
- name: Reload postgresql
|
||||
service: name=postgresql state=reloaded
|
||||
when: postgresql_enabled
|
||||
|
||||
- name: Reload pgpool2
|
||||
service: name=pgpool2 state=reloaded
|
||||
when: pgpool_enabled
|
||||
|
||||
- name: Restart pgpool2
|
||||
service: name=pgpool2 state=reloaded
|
||||
when: pgpool_enabled
|
||||
|
|
|
@ -6,5 +6,35 @@
|
|||
|
||||
- name: Configure pcp
|
||||
template: src=pcp.conf.j2 dest=/etc/pgpool2/pcp.conf owner=root group=postgres mode=0640
|
||||
tags: [ 'postgresql', 'postgres', 'pgpool', 'pcp_conf' ]
|
||||
tags: [ 'postgresql', 'postgres', 'pgpool', 'pcp_conf', 'pgpool_conf' ]
|
||||
|
||||
- name: Install the pgpool configuration file
|
||||
template: src=pgpool.conf.j2 dest=/etc/pgpool2/pgpool.conf owner=root group=postgres mode=0640
|
||||
notify: Restart pgpool2
|
||||
tags: [ 'postgresql', 'postgres', 'pgpool', 'pgpool_conf' ]
|
||||
|
||||
- name: Give access to the remote postgresql clients
|
||||
lineinfile: name=/etc/pgpool2/pool_hba.conf regexp="^host {{ item.0.name }} {{ item.0.user }} {{ item.1 }}.*$" line="host {{ item.0.name }} {{ item.0.user }} {{ item.1 }} md5"
|
||||
with_subelements:
|
||||
- '{{ psql_db_data | default([]) }}'
|
||||
- allowed_hosts
|
||||
when:
|
||||
- psql_db_data is defined
|
||||
- item.1 is defined
|
||||
notify: Reload pgpool2
|
||||
tags: [ 'postgresql', 'postgres', 'pgpool', 'pgpool_conf' ]
|
||||
|
||||
- name: Install the pool_passwd configuration file
|
||||
shell: cd /etc/pgpool2 ; pg_md5 -m -u {{ item.user }} {{ item.pwd }} ; chown root:postgres /etc/pgpool2/pool_passwd ; chmod 660 /etc/pgpool2/pool_passwd
|
||||
with_items: '{{ psql_db_data | default([]) }}'
|
||||
tags: [ 'postgresql', 'postgres', 'pgpool', 'pgpool_conf' ]
|
||||
|
||||
- name: Start and enable pgpool2
|
||||
service: name=pgpool2 state=started enabled=yes
|
||||
when: pgpool_enabled
|
||||
tags: [ 'postgresql', 'postgres', 'pgpool' ]
|
||||
|
||||
- name: Stop and disable pgpool2
|
||||
service: name=pgpool2 state=stopped enabled=no
|
||||
when: not pgpool_enabled
|
||||
tags: [ 'postgresql', 'postgres', 'pgpool' ]
|
||||
|
|
|
@ -0,0 +1,682 @@
|
|||
# ----------------------------
|
||||
# pgPool-II configuration file
|
||||
# ----------------------------
|
||||
#
|
||||
# This file consists of lines of the form:
|
||||
#
|
||||
# name = value
|
||||
#
|
||||
# Whitespace may be used. Comments are introduced with "#" anywhere on a line.
|
||||
# The complete list of parameter names and allowed values can be found in the
|
||||
# pgPool-II documentation.
|
||||
#
|
||||
# This file is read on server startup and when the server receives a SIGHUP
|
||||
# signal. If you edit the file on a running system, you have to SIGHUP the
|
||||
# server for the changes to take effect, or use "pgpool reload". Some
|
||||
# parameters, which are marked below, require a server shutdown and restart to
|
||||
# take effect.
|
||||
#
|
||||
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# CONNECTIONS
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
# - pgpool Connection Settings -
|
||||
|
||||
listen_addresses = '{{ pgpool_listen_addresses }}'
|
||||
# Host name or IP address to listen on:
|
||||
# '*' for all, '' for no TCP/IP connections
|
||||
# (change requires restart)
|
||||
port = {{ pgpool_port }}
|
||||
# Port number
|
||||
# (change requires restart)
|
||||
socket_dir = '/var/run/postgresql'
|
||||
# Unix domain socket path
|
||||
# The Debian package defaults to
|
||||
# /var/run/postgresql
|
||||
# (change requires restart)
|
||||
listen_backlog_multiplier = {{ pgpool_listen_backlog_multiplier }}
|
||||
# Set the backlog parameter of listen(2) to
|
||||
# num_init_children * listen_backlog_multiplier.
|
||||
# (change requires restart)
|
||||
|
||||
# - pgpool Communication Manager Connection Settings -
|
||||
|
||||
pcp_listen_addresses = '{{ pgpool_pcp_listen_addresses }}'
|
||||
# Host name or IP address for pcp process to listen on:
|
||||
# '*' for all, '' for no TCP/IP connections
|
||||
# (change requires restart)
|
||||
pcp_port = {{ pgpool_pcp_port }}
|
||||
# Port number for pcp
|
||||
# (change requires restart)
|
||||
pcp_socket_dir = '/var/run/postgresql'
|
||||
# Unix domain socket path for pcp
|
||||
# The Debian package defaults to
|
||||
# /var/run/postgresql
|
||||
# (change requires restart)
|
||||
|
||||
# - Backend Connection Settings -
|
||||
|
||||
{% for bk in pgpool_backends %}
|
||||
backend_hostname{{ bk.id}} = '{{ bk.hostname }}'
|
||||
backend_port{{ bk.id }} = {{ bk.backend_port }}
|
||||
backend_weight{{ bk.id }} = {{ bk.backend_weight }}
|
||||
backend_data_directory{{ bk.id }} = '{{ bk.backend_data_directory }}'
|
||||
backend_flag{{ bk.id }} = '{{ bk.backend_flag }}'
|
||||
{% endfor %}
|
||||
# - Authentication -
|
||||
|
||||
enable_pool_hba = {{ pgpool_enable_pool_hba }}
|
||||
# Use pool_hba.conf for client authentication
|
||||
pool_passwd = '{{ pgpool_pool_passwd }}'
|
||||
# File name of pool_passwd for md5 authentication.
|
||||
# "" disables pool_passwd.
|
||||
# (change requires restart)
|
||||
authentication_timeout = 60
|
||||
# Delay in seconds to complete client authentication
|
||||
# 0 means no timeout.
|
||||
|
||||
# - SSL Connections -
|
||||
|
||||
ssl = off
|
||||
# Enable SSL support
|
||||
# (change requires restart)
|
||||
#ssl_key = './server.key'
|
||||
# Path to the SSL private key file
|
||||
# (change requires restart)
|
||||
#ssl_cert = './server.cert'
|
||||
# Path to the SSL public certificate file
|
||||
# (change requires restart)
|
||||
#ssl_ca_cert = ''
|
||||
# Path to a single PEM format file
|
||||
# containing CA root certificate(s)
|
||||
# (change requires restart)
|
||||
#ssl_ca_cert_dir = ''
|
||||
# Directory containing CA root certificate(s)
|
||||
# (change requires restart)
|
||||
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# POOLS
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
# - Pool size -
|
||||
|
||||
num_init_children = {{ pgpool_num_init_children }}
|
||||
# Number of pools
|
||||
# (change requires restart)
|
||||
max_pool = {{ pgpool_max_pool }}
|
||||
# Number of connections per pool
|
||||
# (change requires restart)
|
||||
|
||||
# - Life time -
|
||||
|
||||
child_life_time = {{ pgpool_child_life_time }}
|
||||
# Pool exits after being idle for this many seconds
|
||||
child_max_connections = {{ pgpool_child_max_connections }}
|
||||
# Pool exits after receiving that many connections
|
||||
# 0 means no exit
|
||||
connection_life_time = {{ pgpool_connection_life_time }}
|
||||
# Connection to backend closes after being idle for this many seconds
|
||||
# 0 means no close
|
||||
client_idle_limit = {{ pgpool_client_idle_limit }}
|
||||
# Client is disconnected after being idle for that many seconds
|
||||
# (even inside an explicit transactions!)
|
||||
# 0 means no disconnection
|
||||
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# LOGS
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
# - Where to log -
|
||||
|
||||
log_destination = '{{ pgpool_log_destination }}'
|
||||
# Where to log
|
||||
# Valid values are combinations of stderr,
|
||||
# and syslog. Default to stderr.
|
||||
|
||||
# - What to log -
|
||||
|
||||
log_line_prefix = '%t: pid %p: ' # printf-style string to output at beginning of each log line.
|
||||
|
||||
log_connections = {{ pgpool_log_connections }}
|
||||
# Log connections
|
||||
log_hostname = {{ pgpool_log_hostname }}
|
||||
# Hostname will be shown in ps status
|
||||
# and in logs if connections are logged
|
||||
log_statement = {{ pgpool_log_statement }}
|
||||
# Log all statements
|
||||
log_per_node_statement = {{ pgpool_log_per_node_statement }}
|
||||
# Log all statements
|
||||
# with node and backend informations
|
||||
log_standby_delay = 'none'
|
||||
# Log standby delay
|
||||
# Valid values are combinations of always,
|
||||
# if_over_threshold, none
|
||||
|
||||
# - Syslog specific -
|
||||
|
||||
syslog_facility = 'LOCAL0'
|
||||
# Syslog local facility. Default to LOCAL0
|
||||
syslog_ident = 'pgpool'
|
||||
# Syslog program identification string
|
||||
# Default to 'pgpool'
|
||||
|
||||
# - Debug -
|
||||
|
||||
debug_level = {{ pgpool_debug_level }}
|
||||
# Debug message verbosity level
|
||||
# 0 means no message, 1 or more mean verbose
|
||||
|
||||
#log_error_verbosity = default # terse, default, or verbose messages
|
||||
|
||||
#client_min_messages = notice # values in order of decreasing detail:
|
||||
# debug5
|
||||
# debug4
|
||||
# debug3
|
||||
# debug2
|
||||
# debug1
|
||||
# log
|
||||
# notice
|
||||
# warning
|
||||
# error
|
||||
|
||||
#log_min_messages = warning # values in order of decreasing detail:
|
||||
# debug5
|
||||
# debug4
|
||||
# debug3
|
||||
# debug2
|
||||
# debug1
|
||||
# info
|
||||
# notice
|
||||
# warning
|
||||
# error
|
||||
# log
|
||||
# fatal
|
||||
# panic
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# FILE LOCATIONS
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
pid_file_name = '/var/run/postgresql/pgpool.pid'
|
||||
# PID file name
|
||||
# (change requires restart)
|
||||
logdir = '/var/log/postgresql'
|
||||
# Directory of pgPool status file
|
||||
# (change requires restart)
|
||||
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# CONNECTION POOLING
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
connection_cache = on
|
||||
# Activate connection pools
|
||||
# (change requires restart)
|
||||
|
||||
# Semicolon separated list of queries
|
||||
# to be issued at the end of a session
|
||||
# The default is for 8.3 and later
|
||||
reset_query_list = 'ABORT; DISCARD ALL'
|
||||
# The following one is for 8.2 and before
|
||||
#reset_query_list = 'ABORT; RESET ALL; SET SESSION AUTHORIZATION DEFAULT'
|
||||
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# REPLICATION MODE
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
replication_mode = {{ pgpool_replication_mode }}
|
||||
# Activate replication mode
|
||||
# (change requires restart)
|
||||
replicate_select = {{ pgpool_replicate_select }}
|
||||
# Replicate SELECT statements
|
||||
# when in replication mode
|
||||
# replicate_select is higher priority than
|
||||
# load_balance_mode.
|
||||
|
||||
insert_lock = {{ pgpool_insert_lock }}
|
||||
# Automatically locks a dummy row or a table
|
||||
# with INSERT statements to keep SERIAL data
|
||||
# consistency
|
||||
# Without SERIAL, no lock will be issued
|
||||
lobj_lock_table = '{{ pgpool_lobj_lock_table }}'
|
||||
# When rewriting lo_creat command in
|
||||
# replication mode, specify table name to
|
||||
# lock
|
||||
|
||||
# - Degenerate handling -
|
||||
|
||||
replication_stop_on_mismatch = {{ pgpool_replication_stop_on_mismatch }}
|
||||
# On disagreement with the packet kind
|
||||
# sent from backend, degenerate the node
|
||||
# which is most likely "minority"
|
||||
# If off, just force to exit this session
|
||||
|
||||
failover_if_affected_tuples_mismatch = {{ pgpool_failover_if_affected_tuples_mismatch }}
|
||||
# On disagreement with the number of affected
|
||||
# tuples in UPDATE/DELETE queries, then
|
||||
# degenerate the node which is most likely
|
||||
# "minority".
|
||||
# If off, just abort the transaction to
|
||||
# keep the consistency
|
||||
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# LOAD BALANCING MODE
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
load_balance_mode = {{ pgpool_load_balance_mode }}
|
||||
# Activate load balancing mode
|
||||
# (change requires restart)
|
||||
ignore_leading_white_space = {{ pgpool_ignore_leading_white_space }}
|
||||
# Ignore leading white spaces of each query
|
||||
white_function_list = '{{ pgpool_white_function_list }}'
|
||||
# Comma separated list of function names
|
||||
# that don't write to database
|
||||
# Regexp are accepted
|
||||
black_function_list = '{{ pgpool_black_function_list }}'
|
||||
# Comma separated list of function names
|
||||
# that write to database
|
||||
# Regexp are accepted
|
||||
|
||||
database_redirect_preference_list = ''
|
||||
# comma separated list of pairs of database and node id.
|
||||
# example: postgres:primary,mydb[0-4]:1,mydb[5-9]:2'
|
||||
# valid for streaming replicaton mode only.
|
||||
|
||||
app_name_redirect_preference_list = ''
|
||||
# comma separated list of pairs of app name and node id.
|
||||
# example: 'psql:primary,myapp[0-4]:1,myapp[5-9]:standby'
|
||||
# valid for streaming replicaton mode only.
|
||||
allow_sql_comments = {{ pgpool_allow_sql_comments }}
|
||||
# if on, ignore SQL comments when judging if load balance or
|
||||
# query cache is possible.
|
||||
# If off, SQL comments effectively prevent the judgment
|
||||
# (pre 3.4 behavior).
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# MASTER/SLAVE MODE
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
master_slave_mode = off
|
||||
# Activate master/slave mode
|
||||
# (change requires restart)
|
||||
master_slave_sub_mode = 'slony'
|
||||
# Master/slave sub mode
|
||||
# Valid values are combinations slony or
|
||||
# stream. Default is slony.
|
||||
# (change requires restart)
|
||||
|
||||
# - Streaming -
|
||||
|
||||
sr_check_period = 0
|
||||
# Streaming replication check period
|
||||
# Disabled (0) by default
|
||||
sr_check_user = 'nobody'
|
||||
# Streaming replication check user
|
||||
# This is neccessary even if you disable streaming
|
||||
# replication delay check by sr_check_period = 0
|
||||
sr_check_password = ''
|
||||
# Password for streaming replication check user
|
||||
delay_threshold = 0
|
||||
# Threshold before not dispatching query to standby node
|
||||
# Unit is in bytes
|
||||
# Disabled (0) by default
|
||||
|
||||
# - Special commands -
|
||||
|
||||
follow_master_command = ''
|
||||
# Executes this command after master failover
|
||||
# Special values:
|
||||
# %d = node id
|
||||
# %h = host name
|
||||
# %p = port number
|
||||
# %D = database cluster path
|
||||
# %m = new master node id
|
||||
# %H = hostname of the new master node
|
||||
# %M = old master node id
|
||||
# %P = old primary node id
|
||||
# %r = new master port number
|
||||
# %R = new master database cluster path
|
||||
# %% = '%' character
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# HEALTH CHECK
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
health_check_period = 0
|
||||
# Health check period
|
||||
# Disabled (0) by default
|
||||
health_check_timeout = 20
|
||||
# Health check timeout
|
||||
# 0 means no timeout
|
||||
health_check_user = 'nobody'
|
||||
# Health check user
|
||||
health_check_password = ''
|
||||
# Password for health check user
|
||||
health_check_max_retries = 0
|
||||
# Maximum number of times to retry a failed health check before giving up.
|
||||
health_check_retry_delay = 1
|
||||
# Amount of time to wait (in seconds) between retries.
|
||||
connect_timeout = 10000
|
||||
# Timeout value in milliseconds before giving up to connect to backend.
|
||||
# Default is 10000 ms (10 second). Flaky network user may want to increase
|
||||
# the value. 0 means no timeout.
|
||||
# Note that this value is not only used for health check,
|
||||
# but also for ordinary conection to backend.
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# FAILOVER AND FAILBACK
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
failover_command = ''
|
||||
# Executes this command at failover
|
||||
# Special values:
|
||||
# %d = node id
|
||||
# %h = host name
|
||||
# %p = port number
|
||||
# %D = database cluster path
|
||||
# %m = new master node id
|
||||
# %H = hostname of the new master node
|
||||
# %M = old master node id
|
||||
# %P = old primary node id
|
||||
# %r = new master port number
|
||||
# %R = new master database cluster path
|
||||
# %% = '%' character
|
||||
failback_command = ''
|
||||
# Executes this command at failback.
|
||||
# Special values:
|
||||
# %d = node id
|
||||
# %h = host name
|
||||
# %p = port number
|
||||
# %D = database cluster path
|
||||
# %m = new master node id
|
||||
# %H = hostname of the new master node
|
||||
# %M = old master node id
|
||||
# %P = old primary node id
|
||||
# %r = new master port number
|
||||
# %R = new master database cluster path
|
||||
# %% = '%' character
|
||||
|
||||
fail_over_on_backend_error = {{ pgpool_fail_over_on_backend_error }}
|
||||
# Initiates failover when reading/writing to the
|
||||
# backend communication socket fails
|
||||
# If set to off, pgpool will report an
|
||||
# error and disconnect the session.
|
||||
|
||||
search_primary_node_timeout = 10
|
||||
# Timeout in seconds to search for the
|
||||
# primary node when a failover occurs.
|
||||
# 0 means no timeout, keep searching
|
||||
# for a primary node forever.
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# ONLINE RECOVERY
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
recovery_user = 'nobody'
|
||||
# Online recovery user
|
||||
recovery_password = ''
|
||||
# Online recovery password
|
||||
recovery_1st_stage_command = ''
|
||||
# Executes a command in first stage
|
||||
recovery_2nd_stage_command = ''
|
||||
# Executes a command in second stage
|
||||
recovery_timeout = 90
|
||||
# Timeout in seconds to wait for the
|
||||
# recovering node's postmaster to start up
|
||||
# 0 means no wait
|
||||
client_idle_limit_in_recovery = 0
|
||||
# Client is disconnected after being idle
|
||||
# for that many seconds in the second stage
|
||||
# of online recovery
|
||||
# 0 means no disconnection
|
||||
# -1 means immediate disconnection
|
||||
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# WATCHDOG
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
# - Enabling -
|
||||
|
||||
use_watchdog = off
|
||||
# Activates watchdog
|
||||
# (change requires restart)
|
||||
|
||||
# -Connection to up stream servers -
|
||||
|
||||
trusted_servers = ''
|
||||
# trusted server list which are used
|
||||
# to confirm network connection
|
||||
# (hostA,hostB,hostC,...)
|
||||
# (change requires restart)
|
||||
ping_path = '/bin'
|
||||
# ping command path
|
||||
# (change requires restart)
|
||||
|
||||
# - Watchdog communication Settings -
|
||||
|
||||
wd_hostname = ''
|
||||
# Host name or IP address of this watchdog
|
||||
# (change requires restart)
|
||||
wd_port = 9000
|
||||
# port number for watchdog service
|
||||
# (change requires restart)
|
||||
wd_authkey = ''
|
||||
# Authentication key for watchdog communication
|
||||
# (change requires restart)
|
||||
|
||||
# - Virtual IP control Setting -
|
||||
|
||||
delegate_IP = ''
|
||||
# delegate IP address
|
||||
# If this is empty, virtual IP never bring up.
|
||||
# (change requires restart)
|
||||
ifconfig_path = '/sbin'
|
||||
# ifconfig command path
|
||||
# (change requires restart)
|
||||
if_up_cmd = 'ifconfig eth0:0 inet $_IP_$ netmask 255.255.255.0'
|
||||
# startup delegate IP command
|
||||
# (change requires restart)
|
||||
if_down_cmd = 'ifconfig eth0:0 down'
|
||||
# shutdown delegate IP command
|
||||
# (change requires restart)
|
||||
|
||||
arping_path = '/usr/sbin' # arping command path
|
||||
# (change requires restart)
|
||||
|
||||
arping_cmd = 'arping -U $_IP_$ -w 1'
|
||||
# arping command
|
||||
# (change requires restart)
|
||||
|
||||
# - Behaivor on escalation Setting -
|
||||
|
||||
clear_memqcache_on_escalation = on
|
||||
# Clear all the query cache on shared memory
|
||||
# when standby pgpool escalate to active pgpool
|
||||
# (= virtual IP holder).
|
||||
# This should be off if client connects to pgpool
|
||||
# not using virtual IP.
|
||||
# (change requires restart)
|
||||
wd_escalation_command = ''
|
||||
# Executes this command at escalation on new active pgpool.
|
||||
# (change requires restart)
|
||||
|
||||
# - Lifecheck Setting -
|
||||
|
||||
# -- common --
|
||||
|
||||
wd_lifecheck_method = 'heartbeat'
|
||||
# Method of watchdog lifecheck ('heartbeat' or 'query')
|
||||
# (change requires restart)
|
||||
wd_interval = 10
|
||||
# lifecheck interval (sec) > 0
|
||||
# (change requires restart)
|
||||
|
||||
# -- heartbeat mode --
|
||||
|
||||
wd_heartbeat_port = 9694
|
||||
# Port number for receiving heartbeat signal
|
||||
# (change requires restart)
|
||||
wd_heartbeat_keepalive = 2
|
||||
# Interval time of sending heartbeat signal (sec)
|
||||
# (change requires restart)
|
||||
wd_heartbeat_deadtime = 30
|
||||
# Deadtime interval for heartbeat signal (sec)
|
||||
# (change requires restart)
|
||||
heartbeat_destination0 = 'host0_ip1'
|
||||
# Host name or IP address of destination 0
|
||||
# for sending heartbeat signal.
|
||||
# (change requires restart)
|
||||
heartbeat_destination_port0 = 9694
|
||||
# Port number of destination 0 for sending
|
||||
# heartbeat signal. Usually this is the
|
||||
# same as wd_heartbeat_port.
|
||||
# (change requires restart)
|
||||
heartbeat_device0 = ''
|
||||
# Name of NIC device (such like 'eth0')
|
||||
# used for sending/receiving heartbeat
|
||||
# signal to/from destination 0.
|
||||
# This works only when this is not empty
|
||||
# and pgpool has root privilege.
|
||||
# (change requires restart)
|
||||
|
||||
#heartbeat_destination1 = 'host0_ip2'
|
||||
#heartbeat_destination_port1 = 9694
|
||||
#heartbeat_device1 = ''
|
||||
|
||||
# -- query mode --
|
||||
|
||||
wd_life_point = 3
|
||||
# lifecheck retry times
|
||||
# (change requires restart)
|
||||
wd_lifecheck_query = 'SELECT 1'
|
||||
# lifecheck query to pgpool from watchdog
|
||||
# (change requires restart)
|
||||
wd_lifecheck_dbname = 'template1'
|
||||
# Database name connected for lifecheck
|
||||
# (change requires restart)
|
||||
wd_lifecheck_user = 'nobody'
|
||||
# watchdog user monitoring pgpools in lifecheck
|
||||
# (change requires restart)
|
||||
wd_lifecheck_password = ''
|
||||
# Password for watchdog user in lifecheck
|
||||
# (change requires restart)
|
||||
|
||||
# - Other pgpool Connection Settings -
|
||||
|
||||
#other_pgpool_hostname0 = 'host0'
|
||||
# Host name or IP address to connect to for other pgpool 0
|
||||
# (change requires restart)
|
||||
#other_pgpool_port0 = 5432
|
||||
# Port number for othet pgpool 0
|
||||
# (change requires restart)
|
||||
#other_wd_port0 = 9000
|
||||
# Port number for othet watchdog 0
|
||||
# (change requires restart)
|
||||
#other_pgpool_hostname1 = 'host1'
|
||||
#other_pgpool_port1 = 5432
|
||||
#other_wd_port1 = 9000
|
||||
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# OTHERS
|
||||
#------------------------------------------------------------------------------
|
||||
relcache_expire = {{ pgpool_relcache_expire }}
|
||||
# Life time of relation cache in seconds.
|
||||
# 0 means no cache expiration(the default).
|
||||
# The relation cache is used for cache the
|
||||
# query result against PostgreSQL system
|
||||
# catalog to obtain various information
|
||||
# including table structures or if it's a
|
||||
# temporary table or not. The cache is
|
||||
# maintained in a pgpool child local memory
|
||||
# and being kept as long as it survives.
|
||||
# If someone modify the table by using
|
||||
# ALTER TABLE or some such, the relcache is
|
||||
# not consistent anymore.
|
||||
# For this purpose, cache_expiration
|
||||
# controls the life time of the cache.
|
||||
relcache_size = 256
|
||||
# Number of relation cache
|
||||
# entry. If you see frequently:
|
||||
# "pool_search_relcache: cache replacement happend"
|
||||
# in the pgpool log, you might want to increate this number.
|
||||
|
||||
check_temp_table = on
|
||||
# If on, enable temporary table check in SELECT statements.
|
||||
# This initiates queries against system catalog of primary/master
|
||||
# thus increases load of master.
|
||||
# If you are absolutely sure that your system never uses temporary tables
|
||||
# and you want to save access to primary/master, you could turn this off.
|
||||
# Default is on.
|
||||
|
||||
check_unlogged_table = on
|
||||
# If on, enable unlogged table check in SELECT statements.
|
||||
# This initiates queries against system catalog of primary/master
|
||||
# thus increases load of master.
|
||||
# If you are absolutely sure that your system never uses unlogged tables
|
||||
# and you want to save access to primary/master, you could turn this off.
|
||||
# Default is on.
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# IN MEMORY QUERY MEMORY CACHE
|
||||
#------------------------------------------------------------------------------
|
||||
memory_cache_enabled = {{ pgpool_memory_cache_enabled }}
|
||||
# If on, use the memory cache functionality, off by default
|
||||
memqcache_method = '{{ pgpool_memqcache_method }}'
|
||||
# Cache storage method. either 'shmem'(shared memory) or
|
||||
# 'memcached'. 'shmem' by default
|
||||
# (change requires restart)
|
||||
memqcache_memcached_host = '{{ pgpool_memqcache_memcached_host }}'
|
||||
# Memcached host name or IP address. Mandatory if
|
||||
# memqcache_method = 'memcached'.
|
||||
# Defaults to localhost.
|
||||
# (change requires restart)
|
||||
memqcache_memcached_port = {{ pgpool_memqcache_memcached_port }}
|
||||
# Memcached port number. Mondatory if memqcache_method = 'memcached'.
|
||||
# Defaults to 11211.
|
||||
# (change requires restart)
|
||||
memqcache_total_size = 67108864
|
||||
# Total memory size in bytes for storing memory cache.
|
||||
# Mandatory if memqcache_method = 'shmem'.
|
||||
# Defaults to 64MB.
|
||||
# (change requires restart)
|
||||
memqcache_max_num_cache = 1000000
|
||||
# Total number of cache entries. Mandatory
|
||||
# if memqcache_method = 'shmem'.
|
||||
# Each cache entry consumes 48 bytes on shared memory.
|
||||
# Defaults to 1,000,000(45.8MB).
|
||||
# (change requires restart)
|
||||
memqcache_expire = {{ pgpool_memqcache_expire }} # Memory cache entry life time specified in seconds.
|
||||
# 0 means infinite life time. 0 by default.
|
||||
# (change requires restart)
|
||||
memqcache_auto_cache_invalidation = {{ pgpool_memqcache_auto_cache_invalidation }}
|
||||
# If on, invalidation of query cache is triggered by corresponding
|
||||
# DDL/DML/DCL(and memqcache_expire). If off, it is only triggered
|
||||
# by memqcache_expire. on by default.
|
||||
# (change requires restart)
|
||||
memqcache_maxcache = 409600
|
||||
# Maximum SELECT result size in bytes.
|
||||
# Must be smaller than memqcache_cache_block_size. Defaults to 400KB.
|
||||
# (change requires restart)
|
||||
memqcache_cache_block_size = 1048576
|
||||
# Cache block size in bytes. Mandatory if memqcache_method = 'shmem'.
|
||||
# Defaults to 1MB.
|
||||
# (change requires restart)
|
||||
memqcache_oiddir = '/var/log/pgpool/oiddir'
|
||||
# Temporary work directory to record table oids
|
||||
# (change requires restart)
|
||||
white_memqcache_table_list = ''
|
||||
# Comma separated list of table names to memcache
|
||||
# that don't write to database
|
||||
# Regexp are accepted
|
||||
black_memqcache_table_list = ''
|
||||
# Comma separated list of table names not to memcache
|
||||
# that don't write to database
|
||||
# Regexp are accepted
|
Loading…
Reference in New Issue