[Pgpool-general] Parallel SELECT retrieve only half of the results

[Sandeep Thakkar]

I'm still waiting for the response. Also, how do I check that parallel query is 
working?




________________________________
From: Sandeep Thakkar <sandeeptt at yahoo.com>
To: Alessandro Candini <candini at meeo.it>; pgpool-general at pgfoundry.org
Sent: Wed, February 2, 2011 2:37:55 PM
Subject: Re: [Pgpool-general] Parallel SELECT retrieve only half of the results


We have setup one master and one slave which was created by taking base backup. 
We did not setup the slave to be as standby, that means, this server also can 
run write queries. Then we followed the tutorial to setup Parallel Query. We 
enabled Load balancing and Parallel query mode and disabled rest of the modes. 
Did everything what the pgpool tutorial suggests. We did change one thing 
though. We modified dist_def_pgbench.sql because we had only 2 nodes. Here is 
the diff:

<     SELECT CASE WHEN $1 > 0 AND $1 <= 100000 THEN 0
<         WHEN $1 > 100000 AND $1 <= 200000 THEN 1
<         ELSE 2
---
>     SELECT CASE WHEN $1 > 0 AND $1  <= 150000 THEN 0
>         ELSE 1

Now, in the end when I execute "SELECT * FROM pgbench_accounts bench_parallel" 
using all 3 ports (2 dbservers and 1 pgpool), I found that the query using 
pgpool takes a longer time than the two ports of the dbservers. I guess it 
should infact take less time, right? Please let me if I have done anything 
wrong.

Thanks for your help.




________________________________
From: Alessandro Candini <candini at meeo.it>
To: pgpool-general at pgfoundry.org
Sent: Mon, January 31, 2011 5:04:28 PM
Subject: [Pgpool-general] Parallel SELECT retrieve only half of the results

Hi, I have configured pgpool-II-3.0.1 with postgresql-9.0.2 for parallel 
queries:

listen_addresses = '*'
port = 9999
pcp_port = 9898
socket_dir = '/tmp'
pcp_socket_dir = '/tmp'
backend_socket_dir = '/tmp'
pcp_timeout = 10
num_init_children = 32
max_pool = 4
child_life_time = 0
connection_life_time = 0
child_max_connections = 0
client_idle_limit = 0
authentication_timeout = 60
logdir = '/var/log/pgpool'
pid_file_name = '/var/log/pgpool/pgpool.pid'
replication_mode = true
load_balance_mode = false
replication_stop_on_mismatch = false
failover_if_affected_tuples_mismatch = false
replicate_select = true
reset_query_list = 'ABORT; DISCARD ALL'
white_function_list = ''
black_function_list = 'nextval,setval'
print_timestamp = true
master_slave_mode = false
master_slave_sub_mode = 'stream'
delay_threshold = 100
log_standby_delay = 'if_over_threshold'
connection_cache =  true
health_check_timeout = 10
health_check_period = 10
health_check_user = 'postgis'
failover_command = '/usr/lib/pgpool-II/3.0.1/bin/failover.sh %h %H
/tmp/trigger_file0'
failback_command = '/usr/lib/pgpool-II/3.0.1/bin/failback.sh %h %M
/tmp/trigger_file0 %m'
fail_over_on_backend_error = false
insert_lock = true
ignore_leading_white_space = false
log_statement = false
log_per_node_statement = false
log_connections = false
log_hostname = false
parallel_mode = true
enable_query_cache = false
pgpool2_hostname = ''
system_db_hostname = '192.168.0.128'
system_db_port = 5433
system_db_dbname = 'pgpool'
system_db_schema = 'pgpool_catalog'
system_db_user = 'pgpool'
system_db_password = ''
backend_hostname0 = '192.168.0.128'
backend_port0 = 5433
backend_weight0 = 1
backend_data_directory0 = '/home/database/9.0.2/data'
backend_hostname1 = '192.168.0.125'
backend_port1 =  5433
backend_weight1 = 1
backend_data_directory1 = '/home/database/9.0.2/data'
enable_pool_hba = false
recovery_user = 'postgis'
recovery_password = 'gispost'
recovery_1st_stage_command = 'basebackup.sh'
recovery_2nd_stage_command = ''
recovery_timeout = 60
client_idle_limit_in_recovery = 0
lobj_lock_table = 'pgpool_lobj_lock'
ssl = false
debug_level = 0

I have splitted the same db on the two machines 192.168.0.128 (the System DB)
and 192.168.0.125, but when I perform something like "SELECT COUNT(*) FROM
mytable" it retrieves only the data on the system DB...
I expected to obtain the sum of data on the two nodes.
Where am I wrong?

_______________________________________________
Pgpool-general mailing list
Pgpool-general at pgfoundry.org
http://pgfoundry.org/mailman/listinfo/pgpool-general


      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pgfoundry.org/pipermail/pgpool-general/attachments/20110211/5976b781/attachment.html>