Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!feed1.news.erols.com!howland.erols.net!vixen.cso.uiuc.edu!news.indiana.edu!news
From: jfieber@indiana.edu (John Fieber)
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: Databases that runs on FreeBSD
Date: 30 Jan 1997 20:39:05 GMT
Organization: Indiana University, Bloomington
Lines: 29
Sender: jfieber@fallout.campusview.indiana.edu
Message-ID: <5cr0t9$lau@dismay.ucs.indiana.edu>
References: <5bltbc$r8t@kodak.rdcs.kodak.com>
<5c150q$7nn@uriah.heep.sax.de> <5c1q01$koh@nova.dimensional.com>
<5cm5a2$ssn@news.webspan.net>
NNTP-Posting-Host: fallout.campusview.indiana.edu
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Newsreader: knews 0.9.8
To: gpalmer@webspan.net (Gary Palmer)
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:34760
In article <5cm5a2$ssn@news.webspan.net>,
gpalmer@webspan.net (Gary Palmer) writes:
> How well would the different ``free'' databases handle, say, several
> hundred thousand records, and being queried more than once a second?
> I'm curious for an upcoming development I've heard of ...
[disclaimer: these observations are based on a November 1995 vintage
of postgres. Things may be better now.]
As far as size, postgres seems to handle it quite well. I have a
database where the primary tables are around 310,000 records, with an
auxuilary table over 1 million. The performance difference between
the full database and a test databse a tiny fraction the size is
barely detectable in casual use. However, I have not done any
careful timing, nor have I tested it with bunches of clients hitting
it at the same time.
At the time I was evaluating database options, MSQL couldn't even
achive remotely acceptable performance on the test database. I have
not follewd MSQL developments closely, but I gather a fair amount of
progress has been made.
With Postgres95, the killer is the overhead of adding a new client
connection. If you have a lot of clients making a few queries each,
you will have a problem. In that case, you may need to write a thin
"middle end" that pre-forks an appropriate number of backend
processes like Apache does with httpd daemons.
-john