1

I've just setup PostgreSQL (9.3) as a replication to another server. When I started it with service postgresql-9.3 start, it reported [FAILED] without any further explanation. I don't see any thing being reported to /var/lib/pgsql/9.3/pgstartup.log nor /var/lib/pgsql/9.3/data/pg_log. What should I do? (I'm on CentOS)

peterh
  • 4,914
  • 13
  • 29
  • 44
Phuong Nguyen
  • 703
  • 1
  • 11
  • 27
  • Check `/var/log/daemon`. I don't know how your `syslog` defaults, but the normal logging path from within a daemon frequently does not become established until after it's happy with all it's particulars; so early errors often go "elsewhere". Consider: `grep gres /var/log/* /var/log/*/*` – ericx Aug 12 '14 at 17:47
  • Something really wrong with my VPS. Now I can't even SSH. It connects but then immediately disconnects. – Phuong Nguyen Aug 13 '14 at 05:30

1 Answers1

1

You might check to make sure that the system has sufficient kernel settings. I've had problems in the past where the PostgreSQL init script will try to start postmaster and have it silently fail due to insufficient shared memory settings.

Normally you would see helpful messages explaining the settings that need to be changed, but the init script can hide the output. Try starting postmaster manually - the command on my system is: /usr/bin/pg_ctl start

If you see output complaining about kernel settings, then you know that's the culprit. This page can help you: http://www.postgresql.org/docs/9.0/static/kernel-resources.html

Boscoe
  • 564
  • 3
  • 6