Hi,
We met a failure when upgrade Oracle 11gr2 to 12cr1 on HP-UX system when executing "$ORACLE_HOME/perl/bin/perl catctl.pl -l ${_upg_log_dir} catupgrd.sql".
The current component status is as below:
COMP_ID COMP_NAME VERSION STATUS
------------ ------------------------------------ ---------------- -----------
CATALOG Oracle Database Catalog Views 12.1.0.2.0 UPGRADED
CATPROC Oracle Database Packages and Types 12.1.0.2.0 INVALID
OWM Oracle Workspace Manager 12.1.0.2.0 VALID
XDB Oracle XML Database 12.1.0.2.0 INVALID
The ERRORs shown in catupgrd*.log are like below:
04:21:27 SQL> alter index xdb.xdbhi_idx rebuild;
alter index xdb.xdbhi_idx rebuild
*
M
ERROR at line 1:
ORA-01418: specified index does not exist
...
(lots of errors related to xdb.* in catupgrd0.log)
(catupgrd1.log)
04:21:34 SQL>
04:21:34 SQL> Rem Register EM Exprses files
04:21:34 SQL> Rem FIXME: do we really need execemx now ??
04:21:34 SQL> exec prvt_emx.register_files(TRUE);
BEGIN prvt_emx.register_files(TRUE); END;
*
ERROR at line 1:
ORA-04063: package body "SYS.PRVT_REPORT_TAGS" has errors
ORA-06508: PL/SQL: could not find program unit being called:
"SYS.PRVT_REPORT_TAGS"
ORA-06512: at "SYS.PRVT_EMX", line 436
ORA-06512: at line 1
(catupgrd2.log)
04:21:40 SQL>
04:21:40 SQL> Rem Create the automatic SQL Tuning and automatic SPM evolve tasks
04:21:40 SQL> Rem If the tasks already exist (catproc is being re-run), do not error.
04:21:40 SQL> begin
04:21:40 2 sys.dbms_sqltune_internal.i_create_auto_tuning_task;
04:21:40 3 exception
04:21:40 4 when others then
04:21:40 5 if (sqlcode = -13607 OR -- task already exists
04:21:40 6 sqlcode = -65040) -- operation not allowed inside PDB(lrg 7000350)
04:21:40 7 then
04:21:40 8 null;
04:21:40 9 else
04:21:40 10 raise;
04:21:40 11 end if;
04:21:40 12 end;
04:21:40 13 /
begin
*
ERROR at line 1:
ORA-04045: errors during recompilation/revalidation of SYS.DBMS_SQLTUNE
ORA-21700: object does not exist or is marked for delete
ORA-06508: PL/SQL: could not find program unit being called: "SYS.DBMS_SQLTUNE"
ORA-06512: at line 10
In Oracle alert log, errors like below are shown:
Errors in file /opt/oracle/diag/rdbms/nsms/NSMS/trace/NSMS_ora_16187.trc:
ORA-21700: object does not exist or is marked for delete
Fri Jan 27 12:02:11 2017
XDB SGA reset to NULL.
In the trace file (/opt/oracle/diag/rdbms/nsms/NSMS/trace/NSMS_ora_16187.trc):
*** 2017-01-27 12:02:11.147
*** SESSION ID:(1146.37397) 2017-01-27 12:02:11.147
*** CLIENT ID:() 2017-01-27 12:02:11.147
*** SERVICE NAME:(SYS$USERS) 2017-01-27 12:02:11.147
*** MODULE NAME:(catcon(pid=15936)) 2017-01-27 12:02:11.147
*** CLIENT DRIVER:(SQL*PLUS) 2017-01-27 12:02:11.147
*** ACTION NAME:(non-CDB::@catsqlreg.sql) 2017-01-27 12:02:11.147
ERROR in qm_init_sga_pass1 [21700]
*** 2017-01-27 12:02:11.147
dbkedDefDump(): Starting a non-incident diagnostic dump (flags=0x0, level=1, mask=0x0)
----- Error Stack Dump -----
ORA-21700: object does not exist or is marked for delete
----- Current SQL Statement for this session (sql_id=6k7tvcrar8hch) -----
CREATE TABLE registry$sqlpatch (
patch_id NUMBER, -- Patch ID
action VARCHAR2(15), -- APPLY or ROLLBACK
action_time TIMESTAMP, -- Time of action
description VARCHAR2(100), -- Patch description
logfile VARCHAR2(500), -- Location of logfile
status VARCHAR2(15), -- BEGIN, END, SUCCESS, WITH ERRORS
version VARCHAR2(20), -- First 4 digits of version
patch_uid NUMBER, -- Patch UID
flags VARCHAR2(10), -- Flags for this patch
bundle_series VARCHAR2(30), -- PSU, EXA, WINBUNDLE, etc.
bundle_id NUMBER, -- Bundle ID being installed
bundle_data XMLType, -- Bundledata associated with this install
CONSTRAINT registry$sqlpatch_pk
PRIMARY KEY (patch_id, patch_uid, version, action, action_time)
)
...
We met insufficient SYSAUX tablespace issue in the first trial. After enlarging SYSAUX tablespace, the other errors like above were met.
Under current case, what can we do next?
How to fallback to the original Oracle 11gr2? Or how to proceed to complete the upgrade?
What are the possible reasons for the failure?
Thank you for any suggestions!
Iris