Re: optimizing CleanupTempFiles

Tom Lane wrote:
> Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> writes:
> > Tom Lane wrote:
> >> [ blink... ] Doesn't look like that should happen. What is your
> >> test case?

This was simply a CREATE TEMP TABLE ... ON COMMIT DROP. The file stays
in place until checkpoint (either a manually invoked one, or a
shutdown's)

Don't temp tables use this kind of temp file? I admit I didn't check; I
just assumed they did.

> > Hmph, must be because of the patch from last winter to prevent
> > relfilenode reuse until next checkpoint.
>
> Ah. I had misunderstood Alvaro to say that temp files (the kind under
> discussion up to now) were not unlinked immediately; which would be
> pretty strange given that fd.c is underneath md.c.

The test case where I actually verify that fd.c was used was a WITH
SCROLL cursor. The file does go away like I expected in that case, as
soon as the cursor is destroyed (or as soon as the backend is closed).

--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.