Postgresql problem after upgrade F16 to F17

reader69 · #1 9th June 2012, 04:02 PM

Recently updated my Fedora 16x64 to 17 on my laptop. I using the postggresql/php based SQL ledger accounting software.
After upgrade I getting "could not connect to server: No such file or directory Is the server running locally and accepting connections on Unix domain socket "/tmp/.s.PGSQL.5432"?" error message, when I want to use my software.
Nothing has been changed in configuration files, postgresql is running, I able to dump my databases, but unable to connect via http.

Code:

service postgresql status
postgresql.service - PostgreSQL database server
	  Loaded: loaded (/usr/lib/systemd/system/postgresql.service; disabled)
	  Active: active (running) since Sat, 09 Jun 2012 16:45:47 +0200; 39s ago
	 Process: 4468 ExecStart=/usr/bin/pg_ctl start -D ${PGDATA} -s -o -p ${PGPORT} -w -t 300 (code=exited, status=0/SUCCESS)
	 Process: 4461 ExecStartPre=/usr/bin/postgresql-check-db-dir ${PGDATA} (code=exited, status=0/SUCCESS)
	Main PID: 4472 (postgres)
	  CGroup: name=systemd:/system/postgresql.service
		  ├ 4472 /usr/bin/postgres -D /var/lib/pgsql/data -p 5432
		  ├ 4473 postgres: logger process   
		  ├ 4475 postgres: writer process   
		  ├ 4476 postgres: wal writer process   
		  ├ 4477 postgres: autovacuum launcher process   
		  └ 4478 postgres: stats collector process

The /tmp/.s.PGSQL.5432 files is exists:

Code:

srwxrwxrwx 1 postgres postgres 0 Jun  9 16:45 /tmp/.s.PGSQL.5432

in pg_hba.conf there is :

Code:

 local   all     all     trust
host    all         all         127.0.0.1/32          trust

and nothing else.

Please somebody help me to solve this problem !

reader69 · #2 14th June 2012, 11:47 AM

Finally I found the solution:
in /lib/systemd/system/httpd.service file I had to disable the PrivateTemp 'feature'

Code:

PrivateTmp=false

I have to say, this PrivateTmp feature is very annoying.

yewtc · #3 11th July 2012, 10:58 AM

Thanks for this. Just a note to say that one should do a reboot or at least:

sudo kill -s SIGHUP 1

to get systemd to reload it's config. Just restarting httpd doesn't work.

It would have been nice if this could have been flagged up a little better.