Re: pgsql-server: Tablespaces. - Mailing list pgsql-hackers

From Andreas Pflug
Subject Re: pgsql-server: Tablespaces.
Date
Msg-id 40D32A69.8080900@pse-consulting.de
Whole thread Raw
In response to Re: pgsql-server: Tablespaces.  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane wrote:

>Christopher Kings-Lynne <chriskl@familyhealth.com.au> writes:
>  
>
>>Comment on TABLESPACE is impossible, no?  Tablespaces are a global 
>>relation and pg_description isn't.
>>    
>>
>
>Well, it has the same issues as COMMENT ON DATABASE, which we support,
>though crudely.
>
>Perhaps we should think about creating a shared version of
>pg_description so we could have more reasonable support for comments
>on shared objects.  I'm not in a hurry for this but it would be a
>reasonable TODO item.
>  
>

There are more sharing issues with tablespaces (which are already 
supported in pgadmin3, btw :-)
To drop a tablespace, it must be empty, but it can be quite painful to 
find out which objects are populating it. Currently, every database has 
to be queried for pg_class.reltablespace=<mytablespaceoid>. I'd love to 
show tablespace dependency information, which would require some sort of 
global pg_namespace, pg_class and pg_index.

Any thoughts about this?

Regards,
Andreas




pgsql-hackers by date:

Previous
From: Carlos Guzmán Álvarez
Date:
Subject: Timestamp format question
Next
From: Tom Lane
Date:
Subject: Re: Timestamp format question