What is the reason to not include database settings (like search_path)
to database dump created with "pg_dump -C"?
For me, I've created tmp patch for pg_dump to make my system work
(patch for CVS version is included).

---------- Forwarded message ----------
From: Nikolay Samokhvalov <[EMAIL PROTECTED]>
Date: Oct 9, 2006 12:45 PM
Subject: pg_dump VS alter database ... set search_path ...
To: PostgreSQL-general <pgsql-general@postgresql.org>


What is the best practice for following case:
In my database I have a set (~10) of schemas; my database periodically
is being backed up and restored at another machine. I have set up
search_path via "ALTER DATABASE ... SET search_path TO ..." to make
all needed schemas visible to any user who has appropriate rights. The
problem is that I cannot use pg_dumpall and pg_dump DOES NOT dump this
ALTER command, even being executed with "-C" option.

Using additional restoration script with list of schemas seems not the
best solution, because a set of schemas can be changed and I have not
only one database.

Search in mail archives gives me understanding that this issue is
among not resolved ones (there are pros and cons for including such
ALTER in pg_dump-ing).

Is there any common practice for this [probably very frequent] issue?

Best regards,

Best regards,

Attachment: pg_dump_settings82b1.patch
Description: Binary data

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster

Reply via email to