Re: Tablespace limit feature - Mailing list pgsql-general

From Stephen Frost
Subject Re: Tablespace limit feature
Date
Msg-id 20141106224301.GV28859@tamriel.snowman.net
Whole thread Raw
In response to Re: Tablespace limit feature  (Jim Nasby <Jim.Nasby@BlueTreble.com>)
List pgsql-general
* Jim Nasby (Jim.Nasby@BlueTreble.com) wrote:
> On 11/4/14, 2:10 PM, Alvaro Herrera wrote:
> >Alejandro Carrillo escribió:
> >>>
> >>>But This feature is very necessary to control the amount space expensed by a postgresql user or tablespace. This
featurecould be used in PostgreSQL Sharing Hosting, see: Shared Database Hosting - PostgreSQL wiki 
> >Sure.  Rub hits the road when it comes to actually implementing this,
> >though -- the issue of enforcing the limit without creating a bottleneck
> >to table extension is not trivial to overcome, as I recall.
>
> And index extension. And visibility map extension. And, and, and.

Indexes I could agree with.  I'm not sure that the VM would be
worthwhile to include explicitly; as its size is dictated directly by
the number of pages in the heap, we might be able to include it
implicitly.  There are definitely big questions around how to track
space used and how to figure out if we've reached the limit, in an
efficient and correct way.

> It's definitely a non-trivial problem to solve, and unfortunately there are extremely few hosting environment
providersactive in the community to even ask for the kinds of features they need, which means this just doesn't get
muchattention. 

Agreed.  There are also larger problems (ran into yet another case of
"wait, roles aren't database-specific?!" just today..) than the quota
issue for hosting providers.

    Thanks,

        Stephen

Attachment

pgsql-general by date:

Previous
From: Jim Nasby
Date:
Subject: Re: Tablespace limit feature
Next
From: Seth Pollack
Date:
Subject: does pg_restore respect CLUSTER ON in the schema?