How to restart postgresql 8.4 on Debian Squeeze when there are no clusters defined?

Question

How to restart postgresql 8.4 on Debian Squeeze?

/etc/init.d/postgresql stop/start/restart does nothing (no output given).

Solution

From the online documentation...

pg_ctl restart [-w] [-t seconds] [-s] [-D datadir] [-c] [-m s[mart] | f[ast] | i[mmediate] ] [-o options]

pg_ctl is a utility for starting, stopping, or restarting the PostgreSQL backend server (postgres), or displaying the status of a running server. Although the server can be started manually, pg_ctl encapsulates tasks such as redirecting log output and properly detaching from the terminal and process group. It also provides convenient options for controlled shutdown.

Oh, and since I ignored this detail in your answer, but now have reason to question it... the default location of postgresql 8.4 is: /etc/init.d/postgresql-8.4, not /etc/init.d/postgresql as you have in your question.

Comments

[Daniel Vérité]

This is the right command, but it may do nothing if, for example, there is no cluster defined. Could you add the output of pg_lsclusters ? Also you can invoke sh -x /etc/init.d/postgresql start to see the sequence of commands executed inside the script.

[Daniil Ryzhkov]

pg_lsclusters outputs empty table. Looks like this is what I am looking for. How can I define cluster and will it cause problems with existing postgesql tables?

[Daniel Vérité]

No cluster normally means no existing data. A new cluster can be defined with pg_createcluster but it will be empty. If you had a cluster previously, you need to understand what happened to it. Is there data under /var/lib/postgresql/8.4 ? (try du there)

[Daniil Ryzhkov]

No, /var/lib/postgresql/8.4 does not exsist. But for some reason my postgresql server works for 3 months.

[Daniel Vérité]

Assuming the data is not lost, you need to locate it. Look for directories named pg_* and base. Also the postgres log files (normally in /var/log/postgresql) may shed some light on what happened last time it was running.

[Daniil Ryzhkov]

serverfault.com/questions/422727/… — solution. Thanks to @Daniel Vérité making my problem clearer

[Daniil Ryzhkov]

pg_ctl doesn't work in Debian (command not found). They have pg_ctlcluster, but it requires name of the cluster. How to I get the name of my cluster?

[Daniil Ryzhkov]

Also /etc/init.d/postgresql --help actually gives output: Usage: /etc/init.d/postgresql {start|stop|restart|reload|force-reload|status} [version ..]. And yes, I tried /etc/init.d/postgresql restart 8.4

[HopelessN00b]

@Daniil And what do the status and force-reload switches do? I guess you could always force kill the process, but it's probably more constructive to figure out why it's giving you no output when you try to start, stop or restart it.

[Chida]

@Daniil - Can you post the output of $ ps ax | grep postmaster

[Daniil Ryzhkov]

@ HopelessN00b everything except help outputs nothing. @Chida, ok, wait please

[Daniil Ryzhkov]

@HopelessN00b 3814 ttyp2 S+ 0:00 grep postmaster