topic Re: disaster recory Oracle 734 in Operating System - HP-UX

disaster recory Oracle 734

Kurt Beyers. — Tue, 21 May 2002 08:19:38 GMT

Hi,

my environment is the following:

OBII3.5 cell manager on HP-UX 11.0
Oracle 7.3.4 on HP-UX 11.0 with EBU 2.2.0.9
patches OB are up to date

The Oracle is online backed up with OB. The backup and restore of the database are OK, but I've got a problem with the restore of the recovery catalog database.

When I select the option 'restore EBU catalog database before restore of target database', the restore fails for the following reason (an extract of the OB messages):

Starting parsing of command script "/var/opt/omni/tmp/rest_cat"
RESTORE CATALOG = EBU
Ending parsing of command script

File "/var/opt/omni/tmp/rest_cat" line 1:
EBU-2567: string value expected before "EBU"
on 05/21/2002 [10:45:29] [ 5965 : yyparse ]
EBU-2552: Error detected in command script
on 05/21/2002 [10:45:29] [ 5965 : brcyprs ]
[Major] From: ob2ebu.exe@husux008 "orap" Time: 05/21/02 10:45:29
Restore of the Oracle Catalog failed with exit status "2".

The reason for the failure is that the SID name EBU should be between double quotes in the /var/opt/omni/tmp/rest_cat file (RESTORE CATALOG = "EBU" instead of RESTORE CATALOG = EBU). Has anybody else found a workaround for this problem?

When I try to do the restore from the recovery catalog from the command line (as the Oracle 734 user), it fails with the following problems:

Found backup of Catalog "EBU", sequence 2 of day 16-MAY-2002, BFS "EBU823_02"

Starting restore to disk

Number of parallel I/O streams: 1
Disk input/output size: 4096 bytes
Maximum tape I/O size: 1024 bytes
Buffer size per I/O stream: 4096 bytes

Starting BFS "EBU823_02" on 16-MAY-2002 11:37:42
"/u01/app/oracle/product/734/obackup/admin/EBU823_02"
EBU-4304: Can not open Backup File Set "EBU823_02" in mode 1 with buffer size 1024
on 05/16/2002 11:37:42 [ 29839 : britopen ]
BFS in progress "EBU823_02" cancelled on 16-MAY-2002 11:37:42

Ending restore to disk

EBU-2012: Job 1 failed due to tape management error
on 16-MAY-2002 11:37:42 [ 27693 : brcdocatres ]

Catalog Restore job 1 FAILED on 16-MAY-2002 11:37:42

RESTORE job FAILED on 16-MAY-2002 11:37:42

On metalink, the refer to the media management software (=Omniback) for these errors, but they don't give any usefull explanation or solution.

Can anybody put me in the right direction?

Thanks,
Kurt

Re: disaster recory Oracle 734

Steve Steel — Tue, 21 May 2002 10:59:35 GMT

hi

Check your variables

something like

Error: EBU-4304: Can not open Backup File Set "" in mode 1
EBU-4304: Can not open Backup File Set "0000023a88ea8e" in mode 1 with buffer size 65536on 02/14/2001 14:19:11 [ 444 : britopen ]
BFS in progress "0000013a88ea8e" cancelled on 14-FEB-01 14:19
BFS in progress "0000023a88ea8e" cancelled on 14-FEB-01 14:19
Ending restore to disk
EBU-2012: Job 27 failed due to tape management error on 14-FEB-01 14:19 [ 448 : brcdores ]
Database Restore job 27 FAILED on 14-FEB-01 14:19
RESTORE job FAILED on 14-FEB-01 14:19

Solution:
Export OB2APPANME= into the environment ??? be careful with the SID ??? it is case sensitive
export OB2APPNAME=

steve Steel

Re: disaster recory Oracle 734

Kurt Beyers. — Tue, 21 May 2002 11:16:00 GMT

I still get the same error even when I export OB2APPNAME:

Found backup of Catalog "TEST", sequence 1 of day 21-MAY-2002, BFS "TEST828"
WARNING: Catalog is not empty, current contents will be dropped

Starting catalog drop
Issuing Statement: @@@catdrp;
Catalog successfully dropped.
Ended catalog drop

Starting catalog creation
Issuing Statement: @@@catcre6;
Catalog successfully created.
Ended catalog creation

Starting restore to disk

Number of parallel I/O streams: 1
Disk input/output size: 4096 bytes
Maximum tape I/O size: 1024 bytes
Buffer size per I/O stream: 4096 bytes

Starting BFS "TEST828_01" on 21-MAY-2002 14:24:41
"/u01/app/oracle/product/734/obackup/admin/TEST828_01"
EBU-4304: Can not open Backup File Set "TEST828_01" in mode 1 with buffer size 4
on 05/21/2002 14:24:41 [ 11293 : britopen ]
BFS in progress "TEST828_01" cancelled on 21-MAY-2002 14:24:41

Ending restore to disk

EBU-2012: Job 9 failed due to tape management error
on 21-MAY-2002 14:24:41 [ 9262 : brcdocatres ]

Catalog Restore job 9 FAILED on 21-MAY-2002 14:24:41

RESTORE job FAILED on 21-MAY-2002 14:24:41

Kurt

Re: disaster recory Oracle 734

Steve Steel — Tue, 21 May 2002 13:53:48 GMT

Hi

Options
1)Ask oracle

2)This looks like slow networking.
Parameter SmWaitForFirstClient, in the OmniBack global file
needing to be increased to 60.

3)the time specified in the restore
script was wrong. Adjust the time to just before the archivelogs started to backup.

Steve Steel

Re: disaster recory Oracle 734

Kurt Beyers. — Wed, 22 May 2002 04:27:41 GMT

Hi Steve,

1. Oracle has been contacted a week ago, they are still sleeping however.

2. It is not a network issue: dbase, Omniback and backu device are on the same HP9000 server

3. Restore time isn't an issue either.

However, I've made a mistake with exporting the environment variables. OB2APPNAME needs to be the SID of the target database and not the SID of the recovery catalog. When the target SID was exported, the manual restore of the recovery catalog worked (or at least didn't give an error, still need to test of it contains all the necesarry data). Why the restore fails through OB is still an issue.

Kurt