• No results found

Backup Problems on UNIX Systems

At this stage, you should have performed all the verification steps described in the previous sections. After this, proceed as follows:

1. Check your Oracle Server configuration

To check the configuration, log in to the Oracle server system as the user root or as the Oracle user that is identified as described in

“Configuring an Oracle User in Data Protector on UNIX Systems” on page 29. The identified Oracle user and the user root must also be added to Data Protector admin or operator group. Then run the following command:

/opt/omni/lbin/util_oracle8.exe -CHKCONF <ORACLE_SID>

(HP-UX and Solaris systems) or

/usr/omni/bin/util_oracle8.exe -CHKCONF <ORACLE_SID> (other UNIX systems).

In case of an error, the error number is displayed in the form

*RETVAL*<Error_number>.

To get the error description, start the command:

/opt/omni/lbin/omnigetmsg <set_number> <Error_number>

(HP-UX and Solaris systems) or

/usr/omni/bin/omnigetmsg <set_number> <Error_number>

(other UNIX systems).

The *RETVAL*0 indicates successful configuration.

IMPORTANT It is possible to receive a *RETVAL*0 even though the backup still fails.

This can happen due to the fact that the backup owner is not the Oracle user root or the Oracle user that is identified as described in

“Configuring an Oracle User in Data Protector on UNIX Systems” on page 29.

2. Verify Data Protector internal data transfer using the testbar2 utility

Before you run the testbar2 utility, verify that the Cell Manager name is correctly defined on the Oracle Server system. Check the /etc/opt/omni/client/cell_server (HP-UX and Solaris systems)

or /usr/omni/config/cell/cell_server (other UNIX systems) file, which contains the name of the Cell Manager system. Then run the following command:

On HP-UX and Solaris systems:

/opt/omni/bin/testbar2 -type:Oracle8 -appname:<ORACLE_SID>

-bar:<backup_specification_name> -perform:backup On other UNIX systems:

/usr/omni/bin/testbar2 -type:Oracle8 -appname:<ORACLE_SID>

-bar:<backup_specification_name> -perform:backup

Switch to the Data Protector Manager and examine the errors reported by the testbar2 utility by clicking the Details button in the Data Protector Monitor context.

If the messages indicate problems on the Data Protector side of the integration, proceed as follows:

Create an Oracle backup specification to back up to a null device or file. If the backup succeeds, the problem may be related to the backup devices. Refer to the HP OpenView Storage Data Protector

Administrator’s Guide for instructions on troubleshooting devices.

Data Protector reports “Export of the Recovery Catalog Database Failed” when backing up Oracle9i

Problem The following errors are listed in the Data Protector monitor:

EXP-00008: ORACLE error 6550 encountered ORA-06550: line 1, column 13:

PLS-00201: identifier 'SYS.LT_EXPORT_PKG' must be declared ORA-06550: line 1, column 7:

PL/SQL: Statement ignored

EXP-00083: The previous problem occurred when calling SYS.LT_EXPORT_PKG.schema_info_exp

. exporting statistics

Export terminated successfully with warnings.

[Major] From: ob2rman.exe@machine "MAKI" Time: 10/01/01 16:07:53

Export of the Recovery Catalog Database failed.

Action Log in to Oracle9i SQL Plus and grant the execute permission to the Oracle9i LT_EXPORT_PKG as follows (make sure that the user sys has the SYSDBA permission granted beforehand):

sqlplus 'sys/<password>@CDB as sysdba'

SQL> grant execute on sys.lt_export_pkg to public;

Restart the failed backup session.

Data Protector reports “Cannot allocate/attach shared memory”

Problem Backup fails and the following error message is displayed:

Cannot allocate/attach shared memory (IPC Cannot Allocate Shared Memory Segment)

System error: [13] Permission denied) => aborting

Action Set the OB2SHMEM_IPCGLOBAL omnirc option in the /opt/omni/.omnirc file to 1 in order to use the memory windowing properly, and restart the failed backup session. Refer to the HP OpenView Storage Data Protector Administrator’s Guide for details on using omnirc options.

Backup Fails After a Point in Time Restore and Recovery

Problem Backup fails after a Point in time restore and recovery was performed and the following error is displayed:

RMAN-06004: ORACLE error from recovery catalog database:

RMAN-20003: target database incarnation not found in recovery catalog

Action Connect to the target and recovery catalog database using RMAN and reset the database:

rman target <Target_Database_Login> rcvcat

<Recovery_Catalog_Login>

RMAN> RESET DATABASE;

RMAN> exit

Backup of Archive Logs on RAC Cannot be Performed

Problem On RAC, the archive logs are not installed on a NFS mounted disk.

Backup of archive logs cannot be performed.

Action Edit the archive logs backup specification:

• Add an additional allocate channel command for each node.

• Replace the <ORACLE_SID> in the OB2APPNAME parameter with

<GLOBAL_DB_NAME>.

• Add a command to connect to each instance. The connection

parameters should be given as <username>/<passwd>@<INSTANCE>.

For example, if you are using two nodes, the backup specification might look as follows:

run {

allocate channel 'dev_0' type 'sbt_tape'

parms 'ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=<GLOBAL DB NAME>,OB2BARLIST=RAC_arch)' connect

<username>/<passwd>@<INSTANCE 1>;

allocate channel 'dev_2' type 'sbt_tape'

parms 'ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=<GLOBAL DB NAME>,OB2BARLIST=RAC_arch)' connect

<username>/<passwd>@<INSTANCE 2>;

backup

format 'RAC_arch<QU_%s:%t:%p>.dbf' archivelog all;

}