Wrap-around XID's

From: Mike Tamburro <m(dot)tamburro(at)technologue(dot)com>
To: pgsql-advocacy(at)postgresql(dot)org
Subject: Wrap-around XID's
Date: 2005-02-08 21:56:12
Message-ID: 6.1.1.1.0.20050208164549.024995d0@mail.technologue.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-advocacy

Hello,

Two years ago, I lead a smallish project which implemented a PostgreSQL
v7.3 database on the Linux platform. The size of the database is only about
50MB, but all has been well. So, I have become a PostgreSQL advocate,
preaching its value wherever I go.

However, my company is now planning a large scale enterprise database
system with accompanying API's and web UI's. Of course, I proposed we use
PostgreSQL again, but one of our technology partners objected.

The objection is the wrap-around problem with XID's. We read together the
v8.0 docs and saw the work-around solution (section 21.1.3), using the
VACUUM command. However, our technology partner still warned against using
PostgreSQL.

What I need:

1) More detail on the use of XID's by PostgreSQL, so I can better
understand the problem.

2) An explanation that the work-around is sufficient.

3) I see that PostgreSQL has an 8-byte data type (bigint) which would all
but the solve the problem. Are there any plans to offer 8-byte XID's in
future versions of PostgreSQL?

Thank you very much for your help,

-------
Michael Tamburro
Director, Information Technology Department
Technologue, Inc.
www.technologue.com
610-430-0104 x29

Responses

Browse pgsql-advocacy by date

  From Date Subject
Next Message Joshua D. Drake 2005-02-08 22:15:29 Re: Wrap-around XID's
Previous Message Jean-Paul ARGUDO 2005-02-08 18:33:07 Re: Solutions Linux 2005 Paris : debriefing