Excellent - thanks for the fast response - it was an oracle dba that set it up 
initially so that may explain it - 

Thanks very much

-----Original Message-----
From: Tom Lane <t...@sss.pgh.pa.us> 
Sent: Sunday, January 3, 2021 12:27 PM
To: Thomas Flatley <flatl...@outlook.com>
Cc: pgsql-general@lists.postgresql.org
Subject: Re: Max# of tablespaces

Thomas Flatley <flatl...@outlook.com> writes:
> Hello, I've checked the docs but cant seem to find if there is a max # of 
> tablespaces allowed - I've come across a 9.5 env with 1600 tablespaces - they 
> want to double that - Oracle's max is 64k, I'm not particularly worried about 
> hitting a wall, if there is one , outside of maintenance issues - any 
> assistance would be greatly appreciated.

There's no particular hard limit, though you might start to run into 
OID-starvation problems at a billion or so tablespaces.

On the other hand, it's important to realize that a Postgres tablespace doesn't 
really *do* anything.  It's just a separate subdirectory.
The only functional reason to use a tablespace is if you can place it on a 
separate filesystem.  There is certainly value in being able to do that --- but 
I've never heard of systems having more than a few dozen filesystems mounted.  
Hence, the above issue sounds suspiciously like somebody is expecting Postgres 
tablespaces to do something they don't do.

(I suppose if you are working on a system that has limits on the number of 
files per directory, or performance problems with large values of that, then 
you could use tablespaces as a workaround.
But TBH you'd be better off moving onto a more modern platform.)

                        regards, tom lane


Reply via email to