Good news - the restore worked fine with the new column type and the
application worked fine once I'd added a duplicate column of type text. So
at least I have a workable recovery procedure!
Many thanks for your help.
Marion
-----Original Message-----
From: pgsql-admin-owner@postgresql.org
[mailto:pgsql-admin-owner@postgresql.org]On Behalf Of Marion McKelvie
Sent: 07 April 2004 08:44
To: pgsql-admin@postgresql.org
Subject: Re: [ADMIN] restoring large objects
Good idea but I'm not sure if the application will then handle the changed
column type correctly. I'll give it a go and let you know...
-----Original Message-----
From: pgsql-admin-owner@postgresql.org
[mailto:pgsql-admin-owner@postgresql.org]On Behalf Of Tom Lane
Sent: 07 April 2004 07:08
To: Marion McKelvie
Cc: pgsql-admin@postgresql.org
Subject: Re: [ADMIN] restoring large objects
"Marion McKelvie" <marion@streamlet.co.uk> writes:
> Many thanks - you're right, the references are stored (by the application
> using the database) in a column of type text.
Yuck :-(. A possible solution is to dump the schema and data separately
(schema in a plain text dump, data in an -Fc or -Ft dump since plain
text doesn't support dumping LOs). Then manually edit the schema file
to change the column data type to "oid". Then load. I honestly am not
sure that this will work cleanly, but it's worth trying --- and if it
doesn't work, please let me know where it goes wrong.
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings
