Unix domain socket "/var/run/postgresql/.s.PGSQL.5432" missing

Question

I try to run an application where its database on postgresql. It won't work and I'm getting the following error whenever I try to interact with the Postgres service:

could not connect to server: Aucun fichier ou dossier de ce type
 Is the server running locally and accepting
 connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432`

I checked if Postgresql service is running

sudo systemctl status postgresql 

postgresql.service - PostgreSQL RDBMS
 Loaded: loaded (/lib/systemd/system/postgresql.service; enabled; vendor preset: enab
 Active: active (exited) since Thu 2019年04月04日 10:28:27 +01; 58min ago
 Main PID: 1212 (code=exited, status=0/SUCCESS)
 Tasks: 0 (limit: 3522)
 CGroup: /system.slice/postgresql.service
avril 04 10:28:27 ubuntu systemd[1]: Starting PostgreSQL RDBMS...
avril 04 10:28:27 ubuntu systemd[1]: Started PostgreSQL RDBMS.

checked the content of the socket unix connection

ls /var/run/postgresql/

got this empty folder

brahim@ubuntu:/var/log/postgresql$ ls /var/run/postgresql/
10-main.pg_stat_tmp

checked the content of the data directory

ls /var/lib/postgresql/10/main/

it contains the following

 base pg_logical pg_serial pg_subtrans pg_wal
 global pg_multixact pg_snapshots pg_tblspc pg_xact
 pg_commit_ts pg_notify pg_stat pg_twophase 
 postgresql.auto.conf pg_dynshmem pg_replslot pg_stat_tmp 
 PG_VERSION postmaster.opts

the file postgresql.conf has the following configuration

 # - Connection Settings -
 #listen_addresses = 'localhost'
 port = 5432
 max_connections = 100
 unix_socket_directories = '/var/run/postgresql'

I tried to uncomment the listen_addresses instruction and it didn't work, also I've tried with '*' for all, same thing

the postgresql file log has

2019年04月04日 12:17:57.765 +01 [942] LOG: listening on IPv4 address "127.0.0.1", port 5432
2019年04月04日 12:17:57.767 +01 [942] LOG: listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
2019年04月04日 12:17:58.190 +01 [971] LOG: database system was shut down at 2019年04月03日 20:19:54 +01
2019年04月04日 12:17:58.192 +01 [971] PANIC: replication checkpoint has wrong magic 1886413116 instead of 307747550
2019年04月04日 12:17:59.900 +01 [942] LOG: startup process (PID 971) was terminated by signal 6: Aborted
2019年04月04日 12:17:59.900 +01 [942] LOG: aborting startup due to startup process failure
2019年04月04日 12:17:59.902 +01 [942] LOG: database system is shut down

I am running postgresql 10.6 any thoughts

• Brahim Ayad: @a_horse_with_no_name yes, i've tried that, and i get the fallowing #psql: could not connect to server: Connexion refusée #Is the server running on host "localhost" (127.0.0.1) and accepting #TCP/IP connections on port 5432?
• user1822: Looking closer at the postgres log, it seems your data directory is corrupt: "replication checkpoint has wrong magic 1886413116 instead of 307747550" and the subsequent "aborting startup due to startup process failure". I don't know if that can be properly fixed, but one way would be to get rid of the current data directory, create a new one using initdb then restore your backup.
• Brahim Ayad: @dezso >brahim@ubuntu:/etc/postgresql/10/main$ ls -la /var/run/postgresql total 0 drwxrwsr-x 3 postgres postgres 60 avril 4 13:22 . drwxr-xr-x 37 root root 980 avril 4 12:18 .. drwxr-s--- 2 postgres postgres 40 avril 4 12:17 10-main.pg_stat_tmp as i mentioned whatever commands you put in to interact with postgresql service it generates the same error code for example 'psql'
• jjanes: "replication checkpoint has wrong magic" Is this server supposed to be using logical replication? Is it the publisher or the subscriber?
• Brahim Ayad: @a_horse_with_no_name after that i creat a new one with ./initdb -D /usr/local/pgsql/data the creation was done successfully, but i try to start it with the fallowing sudo ./pg_ctl -D /usr/local/pgsql/data -l logfile start but i get permission denied error #pg_ctl: cannot be run as root Please log in (using, e.g., "su") as the (unprivileged) user that will own the server process# i changed the owner of the /usr/local/pgsql/data/ to the usr brahim ls - la /usr/local/pgsql/data/ got this #total 8 drwxr-xr-x 2 brahim brahim 4096 avril 4 13:55 . drwxr-xr-x 3 root root 4096 avril 4 13:55

Answer 1

Try passing the -a option to ls to also show files starting with the period . character (i.e. "hidden" files):

$ ls -a /var/run/postgresql/
. .. .s.PGSQL.5432 .s.PGSQL.5432.lock

Answer 2

Solución que me funcionó en Postgresql 11:

sudo pg_ctlcluster 11 main start

También revisar que tenga bien configurado pg_hba.conf

• jcolebrand: Hi Idexan, any chance you can add some more detail about why you think his is wrong, and how to configure pg_hba.conf to make this go away? I see that you are using pg_ctlcluster but he isn't. (I could be wrong, I don't see it). This site is about Q&A, and having those answers is sure sometimes helpful, instead of just "copy-paste this it'll work".