Re: is PG able to handle a >500 GB Database? - Mailing list pgsql-general

From Florent Guillaume
Subject Re: is PG able to handle a >500 GB Database?
Date
Msg-id 20010122224133.A19397@twin.efge.org
Whole thread Raw
In response to Re: is PG able to handle a >500 GB Database?  ("Martin A. Marques" <martin@math.unl.edu.ar>)
Responses Re: Re: is PG able to handle a >500 GB Database?  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
> > Well, no, that's the point of this whole discussion. If you give it
> > enough time, oids can wrap around, which apparently doesn't affect the
> > functionning of the database but application that depend on unique oids
> > could get hosed.
> >
> > Yes, the documentation is wrong.
>
> Wow. I guess we need to hear what the developers have to say. AFAITIK oids
> were unique. This paints a new panorama for me. :-( (not a very good one).

Of course if you have a reasonnably behaved database, you won't see a
wraparound in your lifetime.

If, however, like someone else on this list, you have an application
that does 200 requests per second, and that each request generates two
updates, that's 1440000 oids consumed per hour, 34560000 per day,
12614400000 per year, oops that's more than 2^32, oids have wrapped
around nearly three times in the year...

Florent


pgsql-general by date:

Previous
From: Nelio Alves Pereira Filho
Date:
Subject: miliseconds
Next
From: Tom Lane
Date:
Subject: Re: optimizing "between" queries