Running PostgreSQL with Supervisord - postgresql

I want to run PostgreSQL 9.1 using Supervisor on Ubuntu 10.04. At the moment, I manually start PostgreSQL using the init script:
/etc/init.d/postgresql start
According to this post: http://nicksergeant.com/using-postgresql-with-supervisor-on-ubuntu-1010/, I need to modify the PostgreSQL config to make it run on TCP port instead of Unix socket, in order to make PostgreSQL work with Supervisor.
I have two questions regarding this approach:
Considering this is more of hack, is there any implication (e.g. security/permissions, performance, etc) of doing this?
Why cannot we just run the same init script postgresql in Supervisor config? Instead, as shown in the link above, it runs postmaster?
UPDATE:
Thanks to the useful suggestions from both answers below, I have setup a script for Supervisor to invoke PostgreSQL directly:
#!/bin/sh
# This script is run by Supervisor to start PostgreSQL 9.1 in foreground mode
if [ -d /var/run/postgresql ]; then
chmod 2775 /var/run/postgresql
else
install -d -m 2775 -o postgres -g postgres /var/run/postgresql
fi
exec su postgres -c "/usr/lib/postgresql/9.1/bin/postgres -D /var/lib/postgresql/9.1/main -c config_file=/etc/postgresql/9.1/main/postgresql.conf"
I also set the config: /etc/postgresql/9.1/main/start.conf to manual so that PostgreSQL does not start automatically on boot (however, it's not clear to me whether this config is loaded). And then I setup the Supervisor config for postgres as:
[program:postgres]
user=root
group=root
command=/usr/local/bin/run_postgresql.sh
autostart=true
autorestart=true
stderr_logfile=/home/www-data/logs/postgres_err.log
stdout_logfile=/home/www-data/logs/postgres_out.log
redirect_stderr=true
stopsignal=QUIT
So now, I can start PostgreSQL in supervisorctl by doing start postgres, which runs fine. However, after I issue stop postgres, although supervisorctl declares postgres is stopped, the server apparently is still running as I can psql into it.
I wonder if this is a Supervisor config issue, or a PostgreSQL issue. Any suggestion welcome!

The blog post is rather badly written. There is no "TCP mode": the post's suggested method will still listen on a Unix socket, just in a different directory. Comments in the post such as "external pid file - not needed for TCP mode" are very misleading.
postmaster is the traditional name for the postgresql executable (to distinguish the master dispatching process from the backend slaves). For some time now there has been no separate executable, and now it is installed as simply "postgres".
Assuming that Supervisor is broadly simliar to the qmail/daemontools supervise scheme, it would be entirely possible (in fact, quite normal) to have it run a script that sets up the directories and environment, and then execs postgres with the requisite arguments (or propagates arguments given to the wrapper script, which would be unusual with supervise but makes more sense when you have a config file to put arguments into).
The way supervise worked (and I'm going to continue to assume "Supervisor" is the same) is to have the supervisor process run a subprocess as specified, and simply relaunch a new subprocess if it exits. This is based on the idea that the process being launched is a long-lived daemon process that only exits when something goes badly wrong, and that simply restarting it is a valid fix. By contrast, init scripts such as in /etc/init.d run the subprocess and detach it, and return control to their caller- if the subprocess exits, nothing special happens, and it must be restarted manually. If you tried to simply run /etc/init.d/postgresql start from supervise, it would continuously keep spawning postgresql daemons, as the return from the init script would be interpreted as the daemon process having exited, when in fact it had been started and detached.

To avoid auto-starting the service with the /etc/init.d scripts, the package for postgresql 9.1 provides a file /etc/postgresql/9.1/main/start.conf that contains:
# Automatic startup configuration
# auto: automatically start/stop the cluster in the init script
# manual: do not start/stop in init scripts, but allow manual startup with
# pg_ctlcluster
# disabled: do not allow manual startup with pg_ctlcluster (this can be easily
# circumvented and is only meant to be a small protection for
# accidents).
auto
This is the file to modify to avoid auto-start as opposed to moving away /etc/init.d/postgresql as the blog post suggests.
Also, changing unix sockets parameters for lack of /var/run/postgresql doesn't look like the best idea, because it's the default for any program linked with libpq, and because there's no difficulty in creating that directory with the proper permissions, just like it's done by the package's start sequence in /usr/share/postgresql-common/init.d-functions:
# create socket directory
if [ -d /var/run/postgresql ]; then
chmod 2775 /var/run/postgresql
else
install -d -m 2775 -o postgres -g postgres /var/run/postgresql
fi
And although the default value shouldn't cause a problem, note that whether postmaster ultimately stays in foreground or forks and runs in background is controlled by the silent_mode parameter in postgresql.conf. Make sure that it is off.

I am trying to make both tomcat and postgres run under supervisor, and found some hints here : https://serverfault.com/questions/425132/controlling-tomcat-with-supervisor
Here is my modified run_postgresql.sh, using bash :
#!/bin/bash
# This script is run by Supervisor to start PostgreSQL 9.1 in foreground mode
function shutdown()
{
echo "Shutting down PostgreSQL"
pkill postgres
}
if [ -d /var/run/postgresql ]; then
chmod 2775 /var/run/postgresql
else
install -d -m 2775 -o postgres -g postgres /var/run/postgresql
fi
# Allow any signal which would kill a process to stop PostgreSQL
trap shutdown HUP INT QUIT ABRT KILL ALRM TERM TSTP
exec sudo -u postgres /usr/lib/postgresql/9.1/bin/postgres -D /var/lib/postgresql/9.1/main --config-file=/etc/postgresql/9.1/main/postgresql.conf
With this script postgresql stops correctly after supervisorctl stop postgres.

Related

Configuration issue Postgres on Ubuntu?

I have installed Postgres 12 on Ubuntu by building it from source and I am facing two issues:
Although I followed the installation manual from Postgrez, every time I restart my computer, my Postgres server stopz and is no longer seen as a running process.
To start it the first time after install, I do this from the terminal:
/usr/local/pgsql/bin/initdb -D /usr/local/pgsql/data
/usr/local/pgsql/bin/pg_ctl -D /usr/local/pgsql/data -l logfile start
After a restart, to start DB again when I run: /usr/local/pgsql/bin/initdb -D /usr/local/pgsql/data, it throws this error:
initdb: error: directory "/usr/local/pgsql/data" exists but is not empty
If you want to create a new database system, either remove or empty
the directory "/usr/local/pgsql/data" or run initdb
with an argument other than "/usr/local/pgsql/data".
Does that mean that every time I start Postgres after a restart, I have to create a new /data directory?
Upon installing Postgres sing pip or pip3, one can just switch user to postgres and run psql to enter postgres, however now I have to run "/usr/local/bin/psql". Please note I have exported all the paths per https://www.postgresql.org/docs/12/installation.html. How can I fix this? Can an alias be set for this?
After a restart, to start DB again when I run:
/usr/local/pgsql/bin/initdb -D /usr/local/pgsql/data, it throws this
error:
Does that mean that every
time I start Postgres after a restart, I have to create a new /data
directory?
No, quite the opposite. You don't need to initdb after the first time, you just need to start. It is your attempt to initdb when you don't need to which is causing the error message. Note that attempting to initdb isn't doing any harm, because it refused to run. It just generates log/console noise.
Upon installing Postgres sing pip or pip3, one can just switch user to
postgres and run psql to enter postgres, however now I have to run
"/usr/local/bin/psql". Please note I have exported all the paths per
https://www.postgresql.org/docs/12/installation.html. How can I fix
this?
I don't know what your first sentence means, as you don't use pip or pip3 to install PostgreSQL (or at least, the docs don't describe doing so) although you might use them to install psycopg2 to enable python to talk to PostgreSQL.
You could use an alias, but it would probably make more sense to edit ~/.bash_profile to set the PATH, as described from the page you linked to under Environment Variables.
You have to register postgreSQL as a service.
run this:
pg_ctl register [-N servicename] [-U username] [-P password] [-D datadir] [-S a[uto] | d[emand] ] [-w] [-t seconds] [-s] [-o options]
Example:
pg_ctl register -N postgresql -U OS_username -P OS_password -D '/etc/postgresql/12/data' -w
More info in the manual: pg_ctl
Notes:
Username and Password is related to the OS, not postgresql
If you have doubts read the manual.
/usr/local/pgsql/bin/pg_ctl start -D '/usr/local/pgsql/data'
Export following in postgres user account's ~/.bashrc:
LD_LIBRARY_PATH=/usr/local/pgsql/lib
export LD_LIBRARY_PATH
PATH=/usr/local/pgsql/bin:$PATH
export PATH

PostgreSQL server fails to start on ArchLinux: FATAL: could not create lockfile »/run/postgresql/.s.PGSQL.5432.lock«

I am quite new in Arch and a total beginner in PostgreSQL, so this may be a very basic question.
I installed postgresql 11.5-4 from extra and pgadmin 4 from AUR, both seem to be running well.
I created a test DB with the following command:
initdb -D /home/lg/test-db
I got the answer:
You can start the db-server using:
pg_ctl -D /home/lg/test-db -l logdatei start
I tried that and got:
pg_ctl -D /home/lg/test-db -l logdatei start
waiting for serer to start.... stopped
pg_ctl: could not start the server
check the log.
The log only says that the lockfile »/run/postgresql/.s.PGSQL.5432.lock« could not be created, because the folder could not be found. Under /run is no folder called "postgresql". I suppose postgresql can not create this folder, because it does not have the permission. Several posts online posts suggest to change the user/owner of the db to sudo, however. Postgresql prevents this, however. When I try any command as sudo, postgresql tells me that this command can't be run as root. There must be some very basic error in my thinking here, but I have not worked it out for 3 hours.
You'll have to remove /run/postgresql from unix_socket_directories in postgresql.conf before starting the server.
Probably You have /var/run symlinked to /run and run is on tmpfs. You should add something like d /run/postgresql 0755 postgres postgres - into /usr/lib/tmpfiles.d/postgresql.conf

Correct way to terminate a postgres replication mode connection

I have the following code that enables logical decoding in PostgreSQL 9.4:
pg_recvlogical -h localhost --slot test_slot --create-slot
pg_recvlogical -h localhost --slot test_slot --start -f -
I spawn a node.js subprocess to run this code and listen for changes however I'm unsure of what the correct procedure is to terminate the connection. I usually just CNTRL+C from the command line or kill the subprocess in code but I always get a pg_recvlogical: unexpected termination of replication stream: error. What is the correct way to terminate this connection?
The PostgreSQL documentation says,
--start
Begin streaming changes from the logical replication slot specified by --slot, continuing until terminated by a signal. If the server side change stream ends with a server shutdown or disconnect, retry in a loop unless --no-loop is specified.
So I assume the standard Linux signals meanings apply. Therefore, SIGINT or SIGQUIT should be acceptable, giving the app time to finish.
What I do is I send SIGINT, then enter a loop, check if /proc/$PID_/stat exists, and after custom timeout, I send it SIGKILL, and wait $PID_ for it.
kill -2 $PID_
for i in `seq 1 8`; do
if [ ! -f /proc/$PID_/stat ] ; then
RETURN_CODE=`wait $PID_`;
break;
fi
sleep 1;
done
if [ -f /proc/$PID_/stat ] ; then
kill -9 $PID_;
RETURN_CODE=`wait $PID_`;
fi
This is Bash, but any decent platform/language will have tools to do it.
Shortened, so there may be a bug :)
I did the following to kill a zombie active replication slot.
connect to the server where postgres is installed using CLI with ssh.
ssh -i <generated_security_key_name/path>.pem <username>#<hostIp>
once connected with the VM in your server. type following command to connect to postgres sql.
sudo -i -u <postgres-username>
you can go with psql which is a terminal-based front-end to PostgreSQL
psql
check the pid against your replication slot.
select * from pg_replication_slots;
Enter the following to kill it.
sudo kill <pid>

Automating pg_basebackup command for recovery using cron and script

I have a 2 node system running pacemaker, corosync and postgresql 9.4. I am carrying out pgsql replication using virtual IP and am able to successfully recover a downed machine using manual commands. Now to automate stuff, I want to run the following commands in a script to get my recovered master back on cluster.
`#su -postgres
$rm -rf /var/lib/pgsql/9.4/data/* //To delete old data files
$pg_basebackup -h 192.XX.XX.XX -U postgres -D /var/lib/pgsql/9.4/data -X stream -P // To recover the latest data from standby PC running latest entries
$rm /var/lib/pgsql/tmp/PGSQL.lock
$exit //exit from postgresql shell
#pcs resource cleanup msPostgresql`
Now when i run these commands as a script, it hangs after the first command itself i.e. su -postgres and the cursor blinks at bash$ syntax without inserting the commands down below.
I want to automate this process using cron but the script itself is not working for me. Can someone help me out here.
Regards
As far as I know "su -postgres" is wrong. You can use either "su postgres", or "sudo -i -u postgres"
Regarding the scripts here you can find tested, working scripts. The one you are interested in is called "initiate_replication.sh" there.

How to run postgres on centos when installed via YUM repo as default daemon user

With a freshly installed version of Postgres 9.2 via yum repository on Centos 6, how do you run postgres as a different user when it is configured to run as 'postgres:postgres' (u:g) out of the box?
In addition to AndrewPK's explanation, I'd like to note that you can also start new PostgreSQL instances as any user by stopping and disabling the system Pg service, then using:
initdb -D /path/to/data/directory
pg_ctl start -D /path/to/data/directory
This won't auto-start the server on boot, though. For that you must integrate into your init system. On CentOS 6 a simple System V-style init script in /etc/init.d/ and a suitable symlink into /etc/rc3.d/ or /etc/rc3.d/ (depending on default runlevel) is sufficient.
If running more than one instance at a time they must be on different ports. Change the port directive in postgresql.conf in the datadir or set it on startup with pg_ctl -o "-p 5433" .... You may also need to override the unix_socket_directories if your user doesn't have write permission to the default socket directory.
pg_ctl
initdb
This is only for a fresh installation (as it pertained to my situation) as it involves blowing away the data dir.
The steps I took to resolve this issue while utilizing the packaged startup scripts for a fresh installation:
Remove the postgres data dir /var/lib/pgsql/9.2/data if you've already gone through the initdb process with the postgres user:group configured as default.
Modify the startup script (/etc/init.d/postgresql-9.2) to replace all instances of postgres:postgres with NEWUSER:NEWGROUP.
Modify the startup script to replace all instances of postgres in any $SU -l postgres lines with the NEWUSER.
run /etc/init.d/postgres initdb to regenerate the cluster using the new username
Make sure any logs created are owned by the new user or remove old logs if error on initdb (the configuration file in my case was found in /var/lib/pgsql/9.2/data/postgresql.conf).
Startup postgres and it should now be running under the new user/group.
I understand this might not be what other people are looking for if they have existing postgres db's and want to restart the server to run as a different user/group combo - this was not my case, and I didn't see an answer posted anywhere for a 'fresh' install utilizing the pre-packaged startup scripts.