From: | nolan(at)celery(dot)tssi(dot)com |
---|---|
To: | tgl(at)sss(dot)pgh(dot)pa(dot)us (Tom Lane) |
Cc: | mail(at)joeconway(dot)com (Joe Conway), pgsql-general(at)postgresql(dot)org (pgsql general list) |
Subject: | Re: dump_all/restore times? |
Date: | 2003-07-16 23:39:12 |
Message-ID: | 20030716233912.26461.qmail@celery.tssi.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
> I have a sneaking suspicion that creation of foreign key constraints may
> be unreasonably inefficient during a restore, too. Have not had a
> chance to check up on it though. Next time you run such a restore,
> could you turn on log_statement and log_timestamp (or log_duration if
> you have it) so we can see which steps take the most time?
No foreign keys, but a lot of fairly large tables (by my usual standards),
several of which are multiply indexed.
It didn't seem to get into the indexing part until the last 2-3 hours,
I think most of that time was just loading data.
I can run the load again, this system was set up so I could develop more
familiarity with FreeBSD and to have a place to run CVS, the data
is just there to have something to test 7.4 with.
I was also planning to set up 7.3.3 on it (probably with a subset of the
data) so that I could do some comparative timings during beta testing.
I will try the restore under 7.3.3 to see if it performs much differently,
though after reading Joe's note I suspect that's probably just how long
it takes on this box.
--
Mike Nolan
From | Date | Subject | |
---|---|---|---|
Next Message | Lynna Landstreet | 2003-07-16 23:39:37 | Unicode database question |
Previous Message | Joe Conway | 2003-07-16 23:39:01 | Re: Where is the physical files of database that I just |