Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Usenet -> c.d.o.server -> Re: Multiple SID's (instances) possible with Oracle 8 (Unix)?

Re: Multiple SID's (instances) possible with Oracle 8 (Unix)?

From: Eric de Redelijkheid <>
Date: Wed, 28 Dec 2005 16:20:24 +0100
Message-ID: <43b2ad37$0$11063$>

Anno Domini 28-12-2005 15:02, sprak aldus:

>Hi guys,
>sorry to bug you, but I've got a (hopefully) simple question:
>We have a client in South-Korea running an Oracle Database. They are
>using Oracle 8 on some sort of Unix (I guess it's Linux), and they are
>using a characterset KO16KSC5601
>We are running Oracle 8 on a W2K-server having the encoding
>Here, with our encoding, everything works fine, Korean characters are
>displayed smoothley. However, when I'm sending them some load-data
>(tab-separated for SQL-Loader), there are always error-messages (I
>could look up the exact errorcode if necessary).
>So, looking at our W2K Oracle-installation, I suggested to create
>another SID on their Unix-system with the encoding WE8ISO8859P1, and
>they should be done.
>However, their DBA said that this would not be possible, he didn't give
>me any reason.
>Now my question to you is:
>Is it true that within Unix (text-only, no graphs) you can only run ONE
>SID with exactly ONE encoding?
>If not, is there anywhere a "howto" or something with which I can slap
>the DBA's ears?
>Or, do you gys have any suggestion how migrating one character-set into
>the other?
>Please be aware that I'm neither a DBA nor an Oracle-specialist.
>TIA & best regards from Cologne, Germany!

Yes, you can have multiple instances with different language-settings and even multiple Oracle-versions running on the same server.

You cannot simply switch or copy the datafiles to another instance with a different language; you would have to use import/export unless there is a way as discribed in Metalink note 225912.1. In short: you would give the command "ALTER DATABASE CHARACTER SET" and the database has a new character set.

However, it would work only if the source character set is a valid subset of the target character set. This is not the case.

Within this Metalink-note, there are links to other notes. Read them if this note applies to you (and if you have access to Metalink).

Furthermore there could be a difference in the NLS_LANG parameter of the client and the database. A character that might appear as a valid korean symbol on the client, is probably a latin character in the database. WE8ISO8859P1 is a latin character set. So the load data you are sending might not make sense to the other database.

Error messages are always usefull. There might be another problem. Received on Wed Dec 28 2005 - 09:20:24 CST

Original text of this message