I don't know if anyone else has had this problem but i found that the
/etc/init.d/postgresql script simply would not start up my databases
after a reboot. though it worked fine just after creating the database
and getting all the config files in the expected places. Was able to
down and up the dbs all day long, if I wished. Then I rebooted and there
was no databases started. Executing /etc/init.d/postgresql start 9.2
resulted in "No database clusters on the system" or suchlike.
I tracked the problem down to the script calling on pg_lsclusters, which
in turn looks for a /var/run/postgresql directory. All well and good,
except that /var/run/postgresql is on a tempfs filesystem and is wiped
out after a reboot. (Q: should that even be on a tempfs?)
The start command in /etc/init.d/postgresql does this check and then
executes the function that will/can create the /var/run/postgresql if
the aforementioned check passes...which it won't. So instead you get a
message that there are no database clusters on the server.
It's a fairly new install of 12.04 and I was wondering if subsequent
updates take care of it or am I going to have to hack the start/stop
script and maintain that over package updates?
Ubuntu 12.04
postgresql 9.2.0.4
x86_64
2 CPUs
Thanks
Tony Dare