I have a dedicated PostgreSQL server and a pgBouncer server. all connections are established through pgBouncer.
I tested the system using Apache JMeter and PHP. The results are weird. The throughput with 500 connections are not bad but when I test using more connections it drops.
This is the test result:
The pgbouncer config:
[databases]
maindb = host=212.212.322.323 port=5432 user=myuser dbname=mydb pool_size=1000 pool_mode=transaction
[pgbouncer]
logfile = /var/log/postgresql/pgbouncer.log
pidfile = /var/run/postgresql/pgbouncer.pid
listen_addr = *
listen_port = 6432
auth_type = trust
auth_file = /etc/pgbouncer/userlist.txt
pool_mode = session
max_client_conn = 1000
default_pool_size = 20
It is known limit of pgbouncer. It uses only one CPU, and doesn't work well under very high number of connections (for smaller number of connects it is fast and effective). There are new projects, that can be used instead pgbouncer for this purpose - Odyssey or pgagroal.
Related
I am using below python code to test failover,
import psycopg2
conn = psycopg2.connect(database="ccs_testdb", host="host1,host2", user="postgre_user",
password="secret", port="5432", target_session_attrs="read-write")
cur = conn.cursor()
cur.execute("select pg_is_in_recovery(), inet_server_addr()")
row = cur.fetchone()
print("recovery =", row[0])
print("server =", row[1])
if host1 goes down then connection is not established with host2 automatically. Does anyone have tried it?
I want to connect to master instance from my application and if master goes down then want to fallback to standby instance which would be host2 in above prog.
with target_session_attrs="any" it worked. Its connecting to next host in list which is standby
So I am running a k8s cluster with 3 pod postgres cluster fronted by a 3 pod pgbouncer cluster. Connecting to that is a batch job with multiple parallel workers which stream data into the database via pgbouncer. If I run 10 of these batch job pods everything works smoothly. If I go up an order of magnitude to 100 job pods, a large portion of them fail to connect to the database with the error got error driver: bad connection. Multiple workers run on the same node (5 worker pods per node) So it's only ~26 pods in the k8s cluster.
What's maddening is I'm not seeing any postgres or pgbouncer error/warning logs in Kibana and their pods aren't failing. Also Prometheus logging shows it to be well under the max connections.
Below are the postgres and pgbouncer configs along with the connection code of the workers.
Relevant Connection Code From Worker:
err = backoff.Retry(func() error {
p.connection, err = gorm.Open(postgres.New(postgres.Config{
DSN: p.postgresUrl,
}), &gorm.Config{Logger: newLogger})
return err
}, backoff.NewExponentialBackOff())
if err != nil {
log.Panic(err)
}
Postgres Config:
postgresql:
parameters:
max_connections = 200
shared_buffers = 4GB
effective_cache_size = 12GB
maintenance_work_mem = 1GB
checkpoint_completion_target = 0.7
wal_buffers = 16MB
default_statistics_target = 100
random_page_cost = 4
effective_io_concurrency = 2
work_mem = 6990kB
min_wal_size = 1GB
max_wal_size = 4GB
max_worker_processes = 6
max_parallel_workers_per_gather = 3
max_parallel_workers = 6
max_parallel_maintenance_workers = 3
PgBouncer Config:
[databases]
* = host=loot port=5432 auth_user=***
[pgbouncer]
listen_port = 5432
listen_addr = *
auth_type = md5
auth_file = /pgconf/users.txt
auth_query = SELECT username, password from pgbouncer.get_auth($1)
pidfile = /tmp/pgbouncer.pid
logfile = /dev/stdout
admin_users = ***
stats_users = ***
default_pool_size = 20
max_client_conn = 600
max_db_connections = 190
min_pool_size = 0
pool_mode = session
reserve_pool_size = 0
reserve_pool_timeout = 5
query_timeout = 0
ignore_startup_parameters = extra_float_digits
Screenshot of Postgres DB Stats
Things I've tried:
Having the jobs connect directly to the cluster IP of the Pgbouncer service to rule out DNS.
Increasing PgBouncer connection pool
I'm not sure what the issue is here since I don't have any errors from the DB side to fix and a basic error message from the job side. Any help would be appreciated and I can add more context if a key piece is missing.
This ended up being an issue of postgres not actually using the configmap I had set. The map was for 200 connections but the actual DB was still at the default of 100.
Not much to learn here other than make sure to check that the configs you set actually propagate to the actual service.
I have pgbouncer.ini file with the below configuration
[databases]
test_db = host=localhost port=5432 dbname=test_db
[pgbouncer]
logfile = /var/log/postgresql/pgbouncer.log
pidfile = /var/run/postgresql/pgbouncer.pid
listen_addr = 0.0.0.0
listen_port = 5433
unix_socket_dir = /var/run/postgresql
auth_type = md5
auth_file = /etc/pgbouncer/userlist.txt
admin_users = postgres
#pool_mode = transaction
pool_mode = session
server_reset_query = RESET ALL;
ignore_startup_parameters = extra_float_digits
max_client_conn = 25000
autodb_idle_timeout = 3600
default_pool_size = 250
max_db_connections = 250
max_user_connections = 250
and I have in my postgresql.conf file
max_connections = 2000
does it effect badly on the performance ? because of max_connections in my postgresql.conf ? or it doesn't mean anything and already the connection handled by the pgbouncer ?
one more question. in pgpouncer configuration, does it right listen_addr = 0.0.0.0 ? or should to be listen_addr = * ?
Is it better to set default_pool_size on PgBouncer equal to the number of CPU cores available on this server?
Shall all of default_pool_size, max_db_connections and max_user_connections to be set with the same value ?
So the idea of using pgbouncer is to pool connections when you can't afford to have a higher number of max_connections in PG itself.
NOTE: Please DO NOT set max_connections to a number like 2000 just like that.
Let's start with an example, if you have a connection limit of 20 and then your app or organization wants to have a 1000 connections at a given time, that is where pooler comes into picture and in this specific case you want the 20 connections to pool that 1000 coming in from the application.
To understand how it actually works let's take a step back and understand what happens when you do not have a connection pooler and only rely on PG config setting for the max connections which in our case is 20.
So when a connection comes in from a client\application etc. the main process of postgresql(PID, i.e. parent ID) spawns a child for that. So each new connection spawns a child process under the main postgres process, like so:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
24379 postgres 20 0 346m 148m 122m R 61.7 7.4 0:46.36 postgres: sysbench sysbench ::1(40120)
24381 postgres 20 0 346m 143m 119m R 62.7 7.1 0:46.14 postgres: sysbench sysbench ::1(40124)
24380 postgres 20 0 338m 137m 121m R 57.7 6.8 0:46.04 postgres: sysbench sysbench ::1(40122)
24382 postgres 20 0 338m 129m 115m R 57.4 6.5 0:46.09 postgres: sysbench sysbench ::1(40126)
So now once a connection request is sent, it is received by the POSTMASTER process and creates a child process at OS level under the main parent process. This connection then has a life span of "unlimited" unless close by the application or you have a time out set for idle connections in postgresql.
Now here comes the situation where it can be a very costly affair to manage the connections with a given compute, if they exceed a certain limit. Meaning n number of connections when served have a given compute cost and after some time the OS won't be able to handle a situation with HUGE connections and will in turn cause contentions at different compute level(i.e. Memory, CPU, I/O).
What if you can use the presently spawned child processes(backends) if they are not doing any work. You will save time on getting the child process(backends) and the additional cost as well(this can be different at times). This is where the pool of connections that are always open help to serve different client requests comes in and is also called pooling.
So basically now you have only n connections available but the pooler can manage n+i number of connections to serve the client requests.
This where pg-bouncer helps to reuse the connections. It can be configured with 3 types of pooling i.e Session pooling, Statement pooling and Transaction pooling. Basically bouncer returns the connection back to the pool once it has done, statement level work or transaction level work etc. Only during session pooling it keeps the connections unless it disconnects.
So basically lower down the number of connections at PG conf file level and tune all settings in the bouncer.ini.
To answer the second part:
one more question. in pgpouncer configuration, does it right listen_addr = 0.0.0.0 ? or should to be listen_addr = * ?
It depends if you have a standalone deployment, server etc.
basically if its on the server itself and you want it to allow connections from everywhere(incoming) use "*" if you want only the local network to be allowed use "127.0.0.0".
For the rest of your questions check this link: pgbouncer docs
I have tried to share a little of what I know, feel free to ask away if anything was unclear or or correct if it was incorrectly mentioned.
I had a PostgreSQL based web application that run into no connection left problem. so I used pgbouncer to solve this problem. It works very much better with pgbouncer but I'm still having some problems with connection limit...
Right now my postgresql.conf file contains:
max_connections = 100
shared_buffers = 128MB
#temp_buffers = 8MB
#max_prepared_transactions = 0
#work_mem = 1MB
#maintenance_work_mem = 16MB
#max_stack_depth = 2MB
#temp_file_limit = -1
#max_files_per_process = 1000
#shared_preload_libraries = ''
and my pgbouncer.ini file:
pool_mode = session
max_client_conn = 100
default_pool_size = 45
reserve_pool_size = 1
It still encounters some connection limit problems at peak time of site.
Can you help me configure this settings to better configuration.
Thanks in advance.
I have the following problem: i am trying to set up a streaming replication scenario with load balancing. I read various tutorials but i cannot find the mistake. The replication does not work. I do not have a "wal sender/receiver process". The archiving works and everytime the master restarts, the archived wal files are copied to the slave. I even do not get any error. And in configuration file(s) everything looks like fine for me, e.g. master:
wal_level = hot_standby
wal_keep_segments = 32
max_wal_senders = 5
max_replication_slots = 5
wal_sender_timeout = 60s
What irritates me the most is that there is no "wal sender process" and there is no error thrown.
Thank you for any idea,
Sven
UPDATE 1: my recovery.conf:
standby_mode = 'on'
primary_conninfo = 'host=arcserver1 port=5432 user=postgres pass=postgres'
restore_command = 'pg_standby /db/pg_archived %f %p >> /var/log/standby.log'
primary_slot_name='standby1'
and my client postgresql.conf contains:
hot_standby = on
I found the solution:i replaced pg_standby with cp, because pg_standby seems to be only for warm standby, not hot standby.