Re: [Slony1-general] Re: Stripping out slony after / before / during pg_restore? - Mailing list pgsql-general

From Christopher Browne
Subject Re: [Slony1-general] Re: Stripping out slony after / before / during pg_restore?
Date
Msg-id 87abdwp357.fsf@dba2.int.libertyrms.com
Whole thread Raw
In response to Re: Stripping out slony after / before / during pg_restore?  (Richard Huxton <dev@archonet.com>)
Responses Re: [Slony1-general] Re: Stripping out slony after / before / during pg_restore?  ("Stuart Bishop" <stuart@stuartbishop.net>)
List pgsql-general
Richard Huxton <dev@archonet.com> writes:
> Glyn Astill wrote:
>> Hi people,
>>
>> I'm setting us up a separate staging / test server and I want to read
>> in a pg_dump of our current origin stripping out all the slony stuff.
>>
>> I was thinking this could serve two purposes a) test out backups
>> restore properly and b) provide us with us with the staging / test
>> server
>>
>> What's the best way to remove all the slony bits?
>
> Well, you can always just drop the slony schema (with a cascade) -
> that should do it.

Not quite.  There are two things that *doesn't* hit:

 a) If there were any tables where a Slony-I key column got added,
    DROP SCHEMA CASCADE doesn't trim that column out, which can
    cause some minor heartburn.  If you don't use TABLE ADD KEY,
    then this isn't an issue.

 b) In versions of Slony-I before 2.0, stripping out the schema on a
    subscriber will leave triggers in a hidden state.

You want to look into UNINSTALL NODE, which does the full cleanup.
--
(format nil "~S@~S" "cbbrowne" "linuxfinances.info")
http://linuxfinances.info/info/languages.html
REALITY is a crutch for people who can't face ITS.

pgsql-general by date:

Previous
From: Marcus Engene
Date:
Subject: Re: regexp_replace() [noindex] thing
Next
From: Reg Me Please
Date:
Subject: Dynamically created cursors vanish in PLPgSQL