Difference between revisions of "PostgreSQL"

From DHVLab

Line 78: Line 78:
recovery_target_timeline = 'latest'
recovery_target_timeline = 'latest'
For further information see: [http://clusterlabs.org/wiki/PgSQL_Replicated_Cluster clusterlabs]

Revision as of 14:54, 8 September 2016

HA PostgreSQL stores the OVirt Engine configuration. It is configured with the following features:

  1. Fail-over of Master
  2. If Master breaks down, RA detects this fault and makes Master to stop, and Slave is promoted to new Master(promote).
  3. Switching between asynchronous and synchronous
  4. If Slave breaks down or LAN has some troubles, the transaction including Write operation will be stopped during the setting of synchronous replication. This means the stop of service. Therefore, RA switches dynamically from the synchronous to the asynchronous replication for prevented from stopping service.
  5. Automated discrimination of data old and new when initial starts
  6. When Pacemaker is started at the same time in the initial starts, RA compares the data of each node using "last xlog replay location" or "last receive replay location" to check which node has the newest data. The node which has the newest data will be Master.

Node 1

su – postgres
mkdir /var/lib/pgsql/pg_archive
cd /var/lib/pgsql/data


listen_addresses = '*'
max_connections = 200
shared_buffers = 32MB
wal_level = hot_standby
synchronous_commit = on 
archive_mode = on    
archive_command = 'cp %p /var/lib/pgsql/pg_archive/%f'
max_wal_senders = 5
wal_keep_segments = 32
replication_timeout = 9000
hot_standby = on 
max_standby_archive_delay = -1
max_standby_streaming_delay = -1
wal_receiver_status_interval = 2
hot_standby_feedback = on 
logging_collector = on
log_filename = 'postgresql-%a.log'
log_truncate_on_rotation = on 
log_rotation_age = 1d  
log_rotation_size = 0
log_timezone = 'Europe/Berlin'
datestyle = 'iso, mdy'
timezone = 'Europe/Berlin'
lc_messages = 'en_US.UTF-8' 
lc_monetary = 'en_US.UTF-8' 
lc_numeric = 'en_US.UTF-8' 
lc_time = 'en_US.UTF-8' 
default_text_search_config = 'pg_catalog.english'
restart_after_crash = off
include '/var/lib/pgsql/tmp/rep_mode.conf'

Add permissions for PostgreSQL replication and database access over network by editing the

local   all             all                                         trust
host    all             all                        trust
host    all             all     ::1/128                             trust
host    all             all                        trust
host    all             all     OVIRTM/32                           md5  
host    all             all     CLUSTERLAN_CLUSTER_N1_IP/32         trust
host    all             all     CLUSTERLAN_CLUSTER_N2_IP/32         trust
host    all             all     CLUSTERLAN_CLUSTER_IP/32            trust
host    replication     all     CLUSTERLAN_CLUSTER_N1_IP/32         trust
host    replication     all     CLUSTERLAN_CLUSTER_N2_IP/32         trust

Node 2

su - postgres
rm -rf /var/lib/pgsql/data/*
pg_basebackup -hCLUSTERLAN_CLUSTER_N1_IP -Upostgres -D/var/lib/pgsql/data -Xstream -P
mkdir /var/lib/pgsql/pg_archive


standby_mode = 'on'
primary_conninfo = 'host=CLUSTERLAN_CLUSTER_N2_IP port=5432 user=postgres
restore_command = 'cp /var/lib/pgsql/pg_archive/%f %p'
recovery_target_timeline = 'latest'

For further information see: clusterlabs