catproc.sq error : PLS-00553: character set name is not recognized

[gopi]

(9i, Solaris 8)
I need the WE8ISO8859P1 char set so I create a DB with:

CREATE DATABASE testdb
.....
CHARACTER SET WE8ISO8859P1
NATIONAL CHARACTER SET AL16UTF16

When running catproc and catrep I get in the spool files a bunch of:
ERROR:
ORA-06550: line 1, column 29:
PLS-00553: character set name is not recognized
ORA-06550: line 0, column 0:
PL/SQL: Compilation unit analysis terminated

My UNIX environment has:
env | grep ORA_NLS33
ORA_NLS33=/oracle/product/9.0.1/ocommon/nls/admin/data
env | grep NLS_LANG
NLS_LANG=AMERICAN_AMERICA.WE8ISO8859P1

Why am I getting the 533 when trying to create the catalog?

[Shestakov]

I don't absolutly sure about this respond.

AL16UTF16 support ASCII coding in second byte of 2-bytes word.
ASCII --> second byte 0000-007F
WE8ISO8859P1 use first byte as ASCII code
WE8ISO8859P1 --> first byte 00-7F

Probably AL16UTF16 and WE8ISO8859P1 incompatible.

U may use UTF8, because:
ASCII(UTF8) --> first byte 00-7F
WE8ISO8859P1 --> first byte 00-7F

[gopi]

I really thought that the incompatibility you mentioned would be the fix - it sounded correct - but did not work.

I opened an iTar.


*** Perhaps *** right now, the only way to achieve WE8ISO8859P1 is by creating the dB with 7 bit ASCII and then using Oracle's procedures to change the character set.

The problem with changing the char set after dB creation is that imports need STATISTICS=none. However, Oracle *should* have this freakin' working correctly!!!!

Thanks Shestakov!

[gopi]

**** Final Solution ***
After jumping through many hoops for Oracle support and sending alert, trace and spool files, the fix is to RECONNECT BEFORE RUNNING CATPROC. It worked.

***** From Oracle Support *****
Its possible you are encountering a known internal bug (unpublished). Please try the following workaround:

Reconnect after Startup.

For example,
1. connect / as sysdba
2. startup
3. connect / as sysdba
4. run catproc.sql