Re: [HACKERS] pg_dump & performance degradation

Don Baccus writes:

> At 01:06 PM 8/1/00 +1000, Philip Warner wrote:
>
> >I agree; it's definitely a non-critical feature. But then, it is only 80
> >lines of code in one place (including 28 non-code lines). I am not totally
> >happy with the results it produces, so I have no objection to removing it
> >all. I just need some more general feedback...
>
> Have you tried pg_dump on a multi-processor machine, which most serious
> database-backed websites run on these days?   Do you see the same performance
> degradation?  My site runs on a dual P450 with RAID 1 LVD disks, and cost
> me exactly $2100 to build (would've been less if I'd laid off the extra
> cooling fans!)

Hi Don.

I'm the one having the problem. The machine in question is a 4-CPU Xeon
with 4GB RAM running Linux and Postgres6.5. It's not a web application
(although AOLserver is running on the same machine). The table has more
than 10 million rows, but I don't know exactly how many.

In any case, I tried Philip's throttle code, ported to pg_dump for 6.5 but
it doesn't seem to help. It seems that the backend doing the COPY is taking
up all the I/O. If there's no way around that I will probably just have to
schedule a couple of hours of downtime to dump the data.

I would like to extend my thanks to Philip and everyone else for their help
and suggestions.

Brian
--
Brian Baquiran <brianb@edsamail.com>
http://www.baquiran.com/  AIM: bbaquiran
Work: +63(2)7182222       Home: +63(2) 9227123

I'm smarter than average. Therefore, average, to me, seems kind of stupid.
People weren't purposely being stupid. It just came naturally.
                              -- Bruce "Tog" Toganazzini