From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | "Jan Cruz" <malebug(at)gmail(dot)com> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: suggestion |
Date: | 2006-02-24 04:45:47 |
Message-ID: | 29115.1140756347@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
"Jan Cruz" <malebug(at)gmail(dot)com> writes:
> On 2/24/06, Michael Glaesemann <grzm(at)myrealbox(dot)com> wrote:
>> If this doesn't do what you want, can you give a bit more
>> explanation? Also, what are you trying to do with this dump file?
> I also want an option that would exclude "CREATE INDEX " whenever
> a schema is being dump
> The reason for this is that whenever I tried to migrate database
> whenever I restore a schema with indexes and then
> restore the data separately it took more or less 24 hours instead of
> the usual 1 hour more or less.
Restoring schema and data separately is guaranteed to be less efficient
than restoring a combined dump. An option to omit indexes from the
schema dump will not fix this. Indeed it will arguably make things
worse --- in the first place there are severe performance issues
associated with unindexed foreign-key checks, and in the second place
there is the foot-gun problem that you might forget to re-add the
indexes at all.
I think the right question to ask here is "why are you so intent on
using separate schema/data restores?" That's not the recommended way
to go about things, and it never will be.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Jan Cruz | 2006-02-24 05:23:44 | Re: suggestion |
Previous Message | Tom Lane | 2006-02-24 04:32:05 | Re: fsutil ideas |