[pgpool-general: 5759] Load-Balancing not working
rainer at ultra-secure.de
rainer at ultra-secure.de
Wed Oct 4 00:14:04 JST 2017
Hi,
I run pgPool-II-96 from the postgresql.org repository.
We use streaming replication, but it's managed outside of pgpool, I
don't want automatic failovers, I don't want pgpool to interfere into
the replication etc.
Just distribute read-queries among the master and the slave.
I have the following configuration:
pid_file_name = '/var/run/pgpool-II-96/pgpool.pid'
listen_addresses = '*'
port = 5432
socket_dir = '/tmp'
pcp_listen_addresses = '*'
pcp_port = 9898
pcp_socket_dir = '/tmp'
backend_hostname0 = '192.168.1.10'
backend_hostname1 = '192.168.1.11'
backend_weight1 = '1'
backend_weight0 = '1'
backend_port0 = '5432'
backend_port1 = '5432'
replication_mode = off
load_balance_mode = on
master_slave_mode = on
master_slave_sub_mode = 'stream'
num_init_children = 601
sr_check_period = 20
sr_check_user = postgres
sr_check_password = the_password
delay_threshold = 60
log_standby_delay = if_over_threshold
However, when I start it and clients start using it, I get:
Oct 3 16:01:12 pgpool-host01-test pgpool: 2017-10-03 16:01:12: pid
3408: LOG: Setting up socket for 0.0.0.0:5432
Oct 3 16:01:12 pgpool-host01-test pgpool: 2017-10-03 16:01:12: pid
3408: LOG: Setting up socket for :::5432
Oct 3 16:01:12 pgpool-host01-test pgpool: 2017-10-03 16:01:12: pid
3408: LOG: pgpool-II successfully started. version 3.6.3 (subaruboshi)
Oct 3 16:01:12 pgpool-host01-test pgpool: 2017-10-03 16:01:12: pid
3408: LOG: find_primary_node: checking backend no 0
Oct 3 16:01:12 pgpool-host01-test pgpool: 2017-10-03 16:01:12: pid
3408: LOG: find_primary_node: primary node id is 0
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
3976: ERROR: unable to read data from frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
3976: DETAIL: EOF encountered with frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
3992: ERROR: unable to read data from frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
3992: DETAIL: EOF encountered with frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
4010: ERROR: unable to read data from frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
4010: DETAIL: EOF encountered with frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
4011: ERROR: unable to read data from frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
4011: DETAIL: EOF encountered with frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
3876: ERROR: unable to read data from frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
3876: DETAIL: EOF encountered with frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
4012: ERROR: unable to read data from frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
4012: DETAIL: EOF encountered with frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
4009: ERROR: unable to read data from frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
4009: DETAIL: EOF encountered with frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
3775: ERROR: unable to read data from frontend
Oct 3 16:03:12 pgpool-host01-test pgpool: 2017-10-03 16:03:12: pid
3775: DETAIL: EOF encountered with frontend
Oct 3 16:07:00 pgpool-host01-test pgpool: 2017-10-03 16:07:00: pid
3801: WARNING: packet kind of backend 0 ['E'] does not match with
master/majority nodes packet kind ['2']
Oct 3 16:07:00 pgpool-host01-test pgpool: 2017-10-03 16:07:00: pid
3801: FATAL: failed to read kind from backend
Oct 3 16:07:00 pgpool-host01-test pgpool: 2017-10-03 16:07:00: pid
3801: DETAIL: kind mismatch among backends. Possible last query was:
"SELECT TRIGGER_NAME, TRIGGER_GROUP, NEXT_FIRE_TIME, PRIORITY FROM
QUARTZ_TRIGGERS WHERE SCHED_NAME =
'PersistedQuartzSchedulerEngineInstance' AND TRIGGER_STATE = $1 AND
NEXT_FIRE_TIME <= $2 AND (MISFIRE_INSTR = -1 OR (MISFIRE_INSTR != -1 AND
NEXT_FIRE_TIME >= $3)) ORDER BY NEXT_FIRE_TIME ASC, PRIORITY DESC" kind
details are: 0[E: prepared statement "S_1" does not exist] 1[2]
Oct 3 16:07:00 pgpool-host01-test pgpool: 2017-10-03 16:07:00: pid
3801: HINT: check data consistency among db nodes
Oct 3 16:07:00 pgpool-host01-test pgpool: 2017-10-03 16:07:00: pid
3408: LOG: child process with pid: 3801 exits with status 512
Oct 3 16:07:00 pgpool-host01-test pgpool: 2017-10-03 16:07:00: pid
3408: LOG: fork a new child process with pid: 4056
Oct 3 16:07:27 pgpool-host01-test pgpool: 2017-10-03 16:07:27: pid
4013: LOG: forked new pcp worker, pid=4062 socket=7
Oct 3 16:07:27 pgpool-host01-test pgpool: 2017-10-03 16:07:27: pid
4013: LOG: PCP process with pid: 4062 exit with SUCCESS.
Oct 3 16:07:27 pgpool-host01-test pgpool: 2017-10-03 16:07:27: pid
4013: LOG: PCP process with pid: 4062 exits with status 0
Oct 3 16:07:59 pgpool-host01-test pgpool: 2017-10-03 16:07:59: pid
4002: WARNING: packet kind of backend 0 ['E'] does not match with
master/majority nodes packet kind ['2']
Oct 3 16:07:59 pgpool-host01-test pgpool: 2017-10-03 16:07:59: pid
4002: FATAL: failed to read kind from backend
Oct 3 16:07:59 pgpool-host01-test pgpool: 2017-10-03 16:07:59: pid
4002: DETAIL: kind mismatch among backends. Possible last query was:
"SELECT TRIGGER_NAME, TRIGGER_GROUP, NEXT_FIRE_TIME, PRIORITY FROM
QUARTZ_TRIGGERS WHERE SCHED_NAME =
'PersistedQuartzSchedulerEngineInstance' AND TRIGGER_STATE = $1 AND
NEXT_FIRE_TIME <= $2 AND (MISFIRE_INSTR = -1 OR (MISFIRE_INSTR != -1 AND
NEXT_FIRE_TIME >= $3)) ORDER BY NEXT_FIRE_TIME ASC, PRIORITY DESC" kind
details are: 0[E: prepared statement "S_1" does not exist] 1[2]
Oct 3 16:07:59 pgpool-host01-test pgpool: 2017-10-03 16:07:59: pid
4002: HINT: check data consistency among db nodes
Oct 3 16:07:59 pgpool-host01-test pgpool: 2017-10-03 16:07:59: pid
3408: LOG: child process with pid: 4002 exits with status 512
Oct 3 16:07:59 pgpool-host01-test pgpool: 2017-10-03 16:07:59: pid
3408: LOG: fork a new child process with pid: 4066
Oct 3 16:08:50 pgpool-host01-test pgpool: 2017-10-03 16:08:50: pid
4056: WARNING: packet kind of backend 0 ['E'] does not match with
master/majority nodes packet kind ['2']
Oct 3 16:08:50 pgpool-host01-test pgpool: 2017-10-03 16:08:50: pid
4056: FATAL: failed to read kind from backend
Oct 3 16:08:50 pgpool-host01-test pgpool: 2017-10-03 16:08:50: pid
4056: DETAIL: kind mismatch among backends. Possible last query was:
"SELECT TRIGGER_NAME, TRIGGER_GROUP, NEXT_FIRE_TIME, PRIORITY FROM
QUARTZ_TRIGGERS WHERE SCHED_NAME =
'PersistedQuartzSchedulerEngineInstance' AND TRIGGER_STATE = $1 AND
NEXT_FIRE_TIME <= $2 AND (MISFIRE_INSTR = -1 OR (MISFIRE_INSTR != -1 AND
NEXT_FIRE_TIME >= $3)) ORDER BY NEXT_FIRE_TIME ASC, PRIORITY DESC" kind
details are: 0[E: prepared statement "S_1" does not exist] 1[2]
Oct 3 16:08:50 pgpool-host01-test pgpool: 2017-10-03 16:08:50: pid
4056: HINT: check data consistency among db nodes
Oct 3 16:08:50 pgpool-host01-test pgpool: 2017-10-03 16:08:50: pid
3408: LOG: child process with pid: 4056 exits with status 512
Oct 3 16:08:50 pgpool-host01-test pgpool: 2017-10-03 16:08:50: pid
3408: LOG: fork a new child process with pid: 4072
This might be related to
http://www.pgpool.net/mantisbt/view.php?id=231
I use: pgpool-II-96-3.6.3-1.rhel7.x86_64
The application in question is probably the liferay portal (it hangs on
startup).
What can I do?
Thanks in advance,
Rainer
More information about the pgpool-general
mailing list