CREATE DATABASE ends up hosing other instances. Now what?

From: Christopher Allen <progplusSPAMBEGONE_at_gte.net>
Date: Tue, 17 Aug 1999 01:37:47 GMT
Message-ID: <L13u3.547$MK5.33178@dfiatx1-snr1.gtei.net>

I have been experimenting today with creating a new instance on an existing Oracle server. The server has two instances already running...or did. My initial script for creating the database was working, and I was running it repeatedly, experimenting with various file sizes. In between each run I would trash the .ctl, .dbf, and .log files generated by the CREATE DATABASE command. These were located in a series of directories separate from the files for the other instances, of course.

This was going along swimmingly until I moved those files to trash without shutting down the database first. When I did issue a SHUTDOWN IMMEDIATE command, svrmgrl started processing the command, and never returned another prompt. I realized what I had done, and moved the files out of trash back to their original locations. However, this did not cause svrmgrl to return, and I eventual had to close the window it was running in.

After that, the database won't start at all, meaning that my two existing instances are now unavailable. This is a test machine, so nobody is complaining, but they will be eventually. When I try to connect now, this is what I get:

/---------------------------------------------------------------------------

\
 my_server% svrmgrl

 Oracle Server Manager Release 2.3.4.0.0 - Production  Copyright (c) Oracle Corporation 1994, 1995. All rights reserved.  Oracle7 Server Release 7.3.4.0.1 - Production  PL/SQL Release 2.3.4.0.0 - Production

 SVRMGR> connect internal
 Connected to an idle instance.
 SVRMGR> mount my_instance

     2> ;
 mount my_instance
 *
 ORA-01034: ORACLE not available
 SVRMGR> shutdown immediate
 ORA-01012: not logged on
 SVRMGR> logon system/manager

     2> ;
 logon system/manager
 *
 ORA-01034: ORACLE not available
 SVRMGR>

\---------------------------------------------------------------------------

/

I cannot log on, I cannot start the database. Basically, I'm stuck. I would appreciate all suggestions for getting the existing instances available again.

In case it is useful, here is the script I was using to create the new instance:

---

spool /u01/app/oracle/EMD_TEST/create/crdbEMD_TEST.lst connect internal
startup nomount pfile=/u01/app/oracle/EMD_TEST/pfile/initEMD_TEST_0.ora

create database "EMD_TEST"

    noarchivelog
    maxdatafiles 30
    maxinstances 8
    maxlogfiles 32
    maxlogmembers 4
    maxdatafiles 30
    character set "US7ASCII"
    datafile
 '/u01/app/oracle/EMD_TEST/system01.dbf' size 40M

    logfile

 GROUP 1 ('/u02/oradata/EMD_TEST/redoEMD_TEST01.dbf',


  '/u03/oradata/EMD_TEST/redoEMD_TEST01.dbf',


  '/u04/oradata/EMD_TEST/redoEMD_TEST01.dbf',


  '/u05/oradata/EMD_TEST/redoEMD_TEST01.dbf') size 10M,


 GROUP 2 ('/u03/oradata/EMD_TEST/redoEMD_TEST02.dbf',


  '/u04/oradata/EMD_TEST/redoEMD_TEST02.dbf',


  '/u05/oradata/EMD_TEST/redoEMD_TEST02.dbf',


  '/u02/oradata/EMD_TEST/redoEMD_TEST02.dbf') size 10M,


 GROUP 3 ('/u04/oradata/EMD_TEST/redoEMD_TEST03.dbf',


  '/u05/oradata/EMD_TEST/redoEMD_TEST03.dbf',


  '/u02/oradata/EMD_TEST/redoEMD_TEST03.dbf',


  '/u03/oradata/EMD_TEST/redoEMD_TEST03.dbf') size 10M,


 GROUP 4 ('/u05/oradata/EMD_TEST/redoEMD_TEST04.dbf',


  '/u02/oradata/EMD_TEST/redoEMD_TEST04.dbf',

'/u03/oradata/EMD_TEST/redoEMD_TEST04.dbf',
'/u04/oradata/EMD_TEST/redoEMD_TEST04.dbf') size 10M;

disconnect
spool off

---

Received on Mon Aug 16 1999 - 20:37:47 CDT