Re: pg_restore issues with intarray - Mailing list pgsql-general

From Adrian Klaver
Subject Re: pg_restore issues with intarray
Date
Msg-id 587d9b02-e4c4-c08e-4b70-169d029bd3e7@aklaver.com
Whole thread Raw
In response to RE: pg_restore issues with intarray  (Kevin Brannen <KBrannen@efji.com>)
Responses Re: pg_restore issues with intarray
List pgsql-general
On 9/5/19 4:06 PM, Kevin Brannen wrote:
>> From: Adrian Klaver <adrian.klaver@aklaver.com>
>>
>> On 9/5/19 2:57 PM, Kevin Brannen wrote:
>>> I think I need some help to understand what’s going here because I
>>> can’t figure it out and google isn’t helping.
>>>
>>> This is for Pg 9.6.5. (Yes, we are upgrading to 12.x in a few months,
>>> or so the plan is.) Pg code came from the community and we compiled it
>>> with no changes. This is on Centos 6.7, though I think the OS doesn’t matter.
>>>
>>> We’re calling pg_restore like:
>>>
>>> $PGPATH/pg_restore -jobs=$NCPU --dbname=x .
>>>
>>> FWIW, the backup was created with:
>>>
>>> $PGPATH/pg_dump --clean --create --format=d --jobs=$NCPU --file=$EXP
>>> --dbname=x
>>
>> The options you are adding for --clean, --create only have meaning for plain text dumps. If you want those actions
tooccur on the restore then add them to the pg_restore line. Though if you are going to create a new database it will
inheritobjects from template1(as you found below), assuming you have not set WITH TEMPLATE to something else.
 
>>
> 
> Good point that I'm not doing plain text dumps.
> 
> Are you saying that my problem is that I need "--clean" on the pg_restore?

No, just that if you where expecting the clean to happen on the restore 
you would be disappointed.

> I can try that. The fact that this only happens on a few DBs and not all still
> mystifies me. See below on the template..

My guess is you where restoring into a database with preexisting objects 
because neither create or clean was being done.



-- 
Adrian Klaver
adrian.klaver@aklaver.com



pgsql-general by date:

Previous
From: Sam Gendler
Date:
Subject: Re: Question from someone who is not trained in computer sciences
Next
From: Adrian Klaver
Date:
Subject: Re: pg_restore issues with intarray