"Ansley, Michael" wrote:
>
> >> Maybe we could only put the temp files in a different directory.
> >> They are where performance matters.
> >> If a normal table create takes a few seconds that is not a
> >> real problem.
> Once you have the ability to create tablespaces, you can modify the temp
> table thingy a little to create temporary tables in the temp tablespace. If
> there is no explicit temp tablespace defined, then it defaults to the system
> tablespace (which is where it goes now anyway).
IMHO, It's good idea placing all temporary files (like pg_sort.tmp etc)
into separate subdirectory.
> >> > We could actually
> >> > hash the file names into directory buckets and access them
> >> > that way, and have one directory that old symlinks to the
> >> > hashed files.
> I don't think this is necessary, because if you have a system that requires
> this kind of action, then the administrator can create a temp tablespace
> which is used for all the temporary tables, and spread the rest of the
> tables and indices among the remaining tablespaces.
According to my practice, one byte hash improve mail performance twice
for 5000 mailboxes. It doesn't significantly increase mail performance
for 1000 mailboxes.
I'm hard to belive postgres data directory with 5000 files ;-))
--
Dmitry Samersoff, DM\S
dms@wplus.net http://devnull.wplus.net
* there will come soft rains
