Re: Q: Length of SID name on Unix systems

From: Jeff Gentry <jlgentry_at_worldnet.att.net>
Date: 1997/10/03
Message-ID: <611dq9$r3c_at_bgtnsc02.worldnet.att.net>#1/1


Don,

It is possible that the impression of a four character limit on the ORACLE_SID is based on slightly out-of-date Oracle Flexible Architecture (OFA) documentation. Only a few years ago, OFA recommended a four character uppercase SID for all databases even though most (but not all) systems supported larger SIDs and database names. This may also be reflected in some older Oracle installation software that automatically created databases.

The newest OFA definition I have (September 1995) changed this with the statement "Use a descriptive 8 character or shorter name." I have found that this and other OFA recommendations contribute to manageability, performance, and reliability of Oracle databases.

Regards,

Jeff Gentry                              Carpe Diem!!!
Computer Systems Authority   Systems Consultant
jgentry_at_csac.com                  jlgentry_at_worldnet.att.net

-- 
Steve Adams <steve.adams_at_acslink.net.au> wrote in article
<34333027.287034624_at_newsserver.trl.oz.au>...

> Hi Don,
>
> Internally, Oracle allows 9 bytes for the database name and 16 bytes
> for the instance name. These are null terminated strings, so the
> available length is one less. For the instance name, all characters,
> even beyond 15 are significant for distinguishing instances, but if an
> instance name exceeds 15 characters Oracle gives up and does not
> record its name internally at all. This is fine in a single instance
> configuration, but I would not be confident to try it in an OPS
> environment, although logic says that it should be OK.
>
> Anyone else?
>
> Regards, Steve Adams
> ----------
> On Thu, 02 Oct 1997 03:33:55 GMT, myname_at_myisp.usa (The Kid) wrote:
>
> >We are bringing up Oracle on a SUN Solaris box
> >
> >It looks like the ORACLE_SID is limited to 4 characters. Is that
> >true? That seems REALLY short, especially since we are going to have
> >a number of instances running.
> >
> >Don
>
>
Received on Fri Oct 03 1997 - 00:00:00 CEST

Original text of this message