Re: Table spaces again [was Re: Threaded Sorting] - Mailing list pgsql-hackers

On 7 Oct 2002 at 16:49, Hans-Jürgen Schönig wrote:

> >Mount a directory on a partition. If the data exceeds on that partition, there 
> >would be disk error. Like tablespace getting overflown. I have seen both the 
> >scenarios in action..
> Of course it can be done somehow. However, with tablespaces it is more 
> db-like and you need not be familiar with the operating system itself.
> Just think of a company having several different operating systems 
> (suns, linux, bsd, ...).
> what do you think could be done in this case? my answer would be an 
> abstraction layer called table spaces ...

OK. Point noted. Suspended till next point.

> >>Quotas are handled differently on ever platform (if available).
> >Yeah. But that's sysadmins responsibility not DBA's.
> Maybe many people ARE the sysadmins of their PostgreSQL box ...
> When developing a database with an open mind people should try to see a 
> problem from more than just one perspective.
> Why should anybody just forget about sysdbas???

If DBA is sysadmin, does it make a difference if he learnes about mount/ln or 
table spaces. Yes it does. Table spaces are limited to databases but mount/ln 
is useful for any general purpose sysadmin work.

That answers the last point as well, I guess..

> >>With tablespaces you can assign 30mb to use a, 120mb to user b etc. ...
> >>Table spaces are a nice abstraction layer to the file system.
> >Hmm.. And how does that fit in database metaphor? What practical use is that? I 
> >can't imagine as I am a developer and not a DBA.
> One of our customers did some minor hosting projects with PostgreSQL. 
> That's what he wanted to have because it is a practical issue.
> a. you don't want to have more than one instance per machine.
> b. you want to assign a certain amount of space to a certain user 
> without using quotas. just think of administration tools - tablespaces 
> are as simple as a select.

Agreed. Perfect point and I didn't thought of it.

But it can be done in directory structure as well. Of course it's quite a 
deviation from what one thinks as plain old directory structure. But if this is 
one point where table spaces win, let's borrow that. There is lot of baggage in 
table spaces that can be left out..

Besides AFAIU, tablespaces implements quota using data files which are pre-
allocated. Pre-claiming space/resource  is the evil of everything likes of 
oracle do and runs in exact opposite direction of postgresql philosophy.

If postgresql has to implement quotas on object, it should do without 
preclaiming space.

Besides if postgresql offers quota on per object basis in directory/object 
scheme, I am sure that's far more granular than tablespaces. Choice is good..
> per directory is a first step - a good step and a good idea but 
> tablespaces are a useful invention. just think of hosting companies, 
> hybrid environments, etc ...
> tablespaces or not a devil and sysdbas may be developers ...

It's not about devil. It's about revaluating need once again. Especially at the 
level of tablespace concept in itself.

ByeShridhar

--
Oblivion together does not frighten me, beloved.        -- Thalassa (in Anne 
Mulhall's body), "Return to Tomorrow",           stardate 4770.3.