Re: Improving backend launch time by preloading relcache - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Improving backend launch time by preloading relcache
Date
Msg-id 13761.1012316755@sss.pgh.pa.us
Whole thread Raw
In response to Re: Improving backend launch time by preloading relcache  (Hiroshi Inoue <Inoue@tpf.co.jp>)
List pgsql-hackers
Hiroshi Inoue <Inoue@tpf.co.jp> writes:
> Probably we don't have to keep the relhasindex info in the
> db any longer and we had better keep some info about REINDEX
> in memory local to the backend.

I never did much care for the "change relhasindex" hack.  Why isn't
IsIgnoringSystemIndexes a sufficient solution?  I don't really care
if REINDEX is a little bit slower than it might be, so just turning
off use of *all* system indexes seems like an adequate answer.

>> Hmm ... what that says is that unlinking pg_internal.init in
>> setRelfilenode is the wrong place.

> Possibly. I couldn't find the appropriate place(way) then
> and so #ifdef's are still there.

Okay.  I'll work on that when I commit the patches I have.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Hannu Krosing
Date:
Subject: Re: sequence indexes
Next
From: Tom Lane
Date:
Subject: Re: inline is not ANSI C