Well, I'm working on a database that gets about 5 Million new records a day
and I've had no scale-related problems. I've found that indexing is sort of
expensive for write performance, but I've worked around that issue. I have
had to put a day's data in a separate table, though. I've found PostgresSql
to be rock solid; that's why I'm planning on migrating my smaller mysql
databses to PostgresSql.
I can't really talk about my application, but if you have any large database
questions you want to ask, feel free.
Mike Diehl.
-----Original Message-----
From: Ron Chmara
To: Andrew Evans
Cc: pgsql-general@postgresql.org
Sent: 10/21/00 1:53 AM
Subject: Re: [GENERAL] MySQL -> pgsql
Andrew Evans wrote:
> On Fri, Oct 20, 2000 at 12:43:32AM -0700, Ron Chmara wrote:
> > I'm a bit late on this thread, but I'm currently wranging a large
> > set of migrations:
> > postgreSQL-> Oracle
> Would you mind explaining why your company's migrating a PostgreSQL
> database to Oracle? I'm hoping to go the other direction, replacing
> Oracle with PostgreSQL for a few commercial web sites.
Scalability concerns.
Specifically, about 2 GB of new record data+indicies per month. 2 Gb
of tiny records (it's for data-mining *massive* website logfiles, in
realtime), putting the live records into the
160-million-records-per-table
count.
When we went searching for "large data set" information, we couldn't
find anybody doing this on postgresql, and decided that we didn't
want to risk being first. :-)
Oracle, IMNSHO, sucks rocks for most things, but for scalabilty, the
table-as-file metaphor has real problems once a single table needs
to hit those insane levels.
-Bop
--
Brought to you from iBop the iMac, a MacOS, Win95, Win98, LinuxPPC
machine,
which is currently in MacOS land. Your bopping may vary.