| Lists: | pgsql-hackers |
|---|
| From: | Gilles Darold <gilles(dot)darold(at)dalibo(dot)com> |
|---|---|
| To: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | pg_dump restore error |
| Date: | 2012-10-14 14:47:40 |
| Message-ID: | 507AD08C.5020603@dalibo.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Lists: | pgsql-hackers |
Hello,
Restoring a backup generated with pg_dump/pg_dumpall in plain text
format and the --clean option will report errors if the backup is loaded
in an other or empty database. I mean that the backup file contains all
SQL order to drop the database's objects before recreating them, so if
you load this backup into a new database it will throw errors on each
DROP call complaining that the objects doesn't exists.
This is not very important because everything goes fine but these error
reports can be easily prevented with the addition of IF EXISTS clauses
and this will probably be less confusing. I've attached a patch adding
those IF EXISTS on each DROP and ALTER statements.
Best regards,
--
Gilles Darold
http://dalibo.com - http://dalibo.org
| Attachment | Content-Type | Size |
|---|---|---|
| pg_dump-drop_if_exists-patch.diff | text/x-patch | 13.0 KB |
| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Gilles Darold <gilles(dot)darold(at)dalibo(dot)com> |
| Cc: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: pg_dump restore error |
| Date: | 2012-10-14 16:50:44 |
| Message-ID: | 26917.1350233444@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Lists: | pgsql-hackers |
Gilles Darold <gilles(dot)darold(at)dalibo(dot)com> writes:
> Restoring a backup generated with pg_dump/pg_dumpall in plain text
> format and the --clean option will report errors if the backup is loaded
> in an other or empty database.
This is intentional. What you propose amounts to a fundamental change
in the semantics of --clean, and I don't think that it's really much
of an improvement.
regards, tom lane
| From: | Gilles Darold <gilles(dot)darold(at)dalibo(dot)com> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: pg_dump restore error |
| Date: | 2012-10-14 22:17:46 |
| Message-ID: | 507B3A0A.6060201@dalibo.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Lists: | pgsql-hackers |
Le 14/10/2012 18:50, Tom Lane a écrit :
> Gilles Darold <gilles(dot)darold(at)dalibo(dot)com> writes:
>> Restoring a backup generated with pg_dump/pg_dumpall in plain text
>> format and the --clean option will report errors if the backup is loaded
>> in an other or empty database.
> This is intentional. What you propose amounts to a fundamental change
> in the semantics of --clean, and I don't think that it's really much
> of an improvement.
I'm agree that this is not an improvement, I used to work with it
because I know it's harmless error messages. The request comes from a
client that claims that replaying his backup on a new server was always
producing those errors.
I was just thinking that cleaning if exists was also harmless, but it is
probably better to have error reported in most of the cases.
Regards,
--
Gilles Darold
Administrateur de bases de données
http://dalibo.com - http://dalibo.org