[Pgpool-general] Load-Balancing works... at least for minutes..

Gaetan Allart gaetan at nexylan.com
Tue Mar 30 07:05:56 UTC 2010


Hi everybody,

I'm a new user of PgPool-II and I'm experiencing a strange issue with it.
I have two PostgreSQL servers acting as master/slave. All INSERT/UPDATE queries go to master and SELECT queries are load-balanced over master and slave.

This configuration works pretty fine, at least for minutes...

Upon startup, pgpool correctly load-balances queries over my two servers. But within minutes, queries sent to slave are decreasing... decreasing... decreasing...
Within 30 minutes, my master gets all queries and becomes sick...

ncp_node_info tells me that status of both servers is all right (status = 2). Even more strange, if I ncp_node_detach and re-attache my slave server, it then gets queries back. And once again, within minutes, load falls down without any particular reason.

Here is my config_file :

listen_addresses = '0.0.0.0'
port = 5433
pcp_port = 9898
socket_dir = '/tmp'
pcp_socket_dir = '/tmp'
backend_socket_dir = '/tmp'
pcp_timeout = 10
num_init_children = 64
max_pool = 4
child_life_time = 0
connection_life_time = 0
child_max_connections = 0
client_idle_limit = 0
authentication_timeout = 10
logdir = '/var/run'
pid_file_name = '/var/run/pgpool/pgpool.pid'
replication_mode = false
load_balance_mode = true
replication_stop_on_mismatch = false
replicate_select = false
reset_query_list = 'ABORT; RESET ALL; SET SESSION AUTHORIZATION DEFAULT'
print_timestamp = true
master_slave_mode = true
connection_cache = true
health_check_timeout = 20
health_check_period = 0
health_check_user = 'nobody'
failover_command = ''
failback_command = ''
insert_lock = true
ignore_leading_white_space = true
log_statement = false
log_connections = false
log_hostname = false
parallel_mode = false
enable_query_cache = false
pgpool2_hostname = ''
backend_hostname0 = 'localhost'
backend_port0 = 5432
backend_weight0 = 0.4
backend_hostname1 = 'Ip.Of.My.Server'
backend_port1 = 5432
backend_weight1 = 0.6
enable_pool_hba = true
recovery_user = 'nobody'
recovery_password = ''
recovery_1st_stage_command = ''
recovery_2nd_stage_command = ''
recovery_timeout = 90
client_idle_limit_in_recovery = 0

My pgpool version :

pgpool-II version 2.3.1 (tomiteboshi)


Any idea where I should look at?


Best regards,
Gaëtan






More information about the Pgpool-general mailing list