Difference between traditional Oracle Datbase backup and using Snapshot Oracle Database Backup
Created: 17 Dec 2012 | Updated: 19 Dec 2012 | 16 comments
This issue has been solved. See solution.
hi,
i m using netbackup 7.5
machine physical(v440) oralce-solaris10 client & database oracle10g release2 and virtual machines(Esx sever) oracle-solarisx86_64 master server and media server.
is some one guide me.
Discussion Filed Under:
Comments 16 Comments • Jump to latest comment
can some one to assist me?
Please refer to the Netbackup 7.5 Oracle Admin guide.
Also refer to Oracle's own Backup and Recovery User's Guide for 10gR2.
The above two guides combined should answer the questions, and possible any other ones too.
A few points (look up the above guides for detailed explanations):
- Netbackup uses Oracle's RMAN scripts for backups.
- When a Nbu policy schedule runs, the Nbu server tiggers a RMAN script on the client to be run. The script has codes that backup the Oracle DB and send it to a 3rd party backup software, which happens to be a Nbu server, thus completing the "circle".
- The "traditional" backup method is often referred to as stream-based. The backup occurs while Oracle is still running and serving, and the restore and recovery could still return a consistent db due to the Redo Logs.
- With RMAN backups, Oracle does not need to be put in to "backup mode".
- NetBackup uses RMAN's "proxy" command to achieve snapshot backups. A snaphost is a point in time copy of the volumes that contain the Oracle db. Like stream based backups, snapshot based backups also do not interfere with Oracle's running and serving. If everything is setup correctly, there should effectively be no down time in service.
- Snapshot based Oracle backups are usually used for SAN storage hardware based offhost alternate client backups. The benefit is that the burden of the backup process will be offloaded entirely from the Oracle host, thus "offhost". (Note: This is also possible with some software based storage solutions.)
- According to the guide, snapshot based backups also deduplicates better; if you use Nbu deduplication.
- Generally, use the "traditional" stream based backups unless you have one or both of the above reasons for going for snapshot based backups.
- Nbu snapshot based backups do not support Oracle ASM storage.
Relon thanx for reply.
can u share the documnet that contain the step by step snapshot oracle datbase policy configuration and restoration.
because i configured it in our test environment but not sure it's correct?
All the steps are already in the Nbu Oracle admin guide I linked to. It has all the policy setup steps as well as RMAN examples for snapshot backups.
Additionally, you may also want to refer to the NetBackup Snapshot Client Administrator's Guide for steps in setting up snapshot based backups for specific storage vendor hardware/software.
If you are looking for something like "Step1: Click this button. Step2: Type this command then press enter. Step3:...", there isn't one.
You could try searching the forum for discussions and examples.
I would recommend focusing on the parts in the document that explain and give examples of using the RMAN "proxy" command to assist Nbu in doing snapshot backups, but without also reading and understanding the rest of the document, it will not help much.
For fundamental concepts in backing up Oracle, you will have to begin with Oracle's own Backup and Recovery user guide. It has more information on the workings of the "proxy" command.
You will also have to refer to the NetBackup Snapshot Client Compatibility List to make sure your specific environment is supported to do snapshot backups in Netbackup.
For example, you will find that using EMC SnapView with the UFS or VxFS file systems for offhost alternate client snapshot backup is supported, but doing the same with ZFS file system isn't.
You will also find that if you do not use 3rd party storage solutions outside of Oracle Solaris (including Veritas Storage Foundation), then it is not possible to do offhost alternate client snapshot backups.
The manuals are good.
I used a combination of Snapshot Client Guide and NBU for Oracle manual to assist a customer some time ago with HP EVA array-based snapshot backups.
Best to get snapshots working with 'normal' filesystem backups, then config Oracle Rman backups.
Supporting Storage Foundation and VCS on Unix and Windows as well as NetBackup on Unix and Windows
Handy NBU Links
in my environment oracle-snapshot policy is already functional but how i restore its image. my means its restore via rman, or via command of snapshot.
root# /usr/openv/netbackup/bin/admincmd/bpimagelist -U -client v440 -d 12/14/2012 -e 12/17/2012
Backed Up Expires Files KB C Sched Type On Hold Index Status Policy
---------------- ---------- -------- -------- - ------------ ------- ------------ ------------
12/17/2012 15:24 12/31/2012 5 2 N Full Backup 0 0 v440_snap_oracle_new
12/17/2012 14:25 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
12/17/2012 14:25 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
12/17/2012 14:23 12/31/2012 5 2 N Full Backup 0 0 v440_snap_oracle_new
12/17/2012 13:25 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
12/17/2012 13:24 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
12/17/2012 13:22 12/31/2012 5 2 N Full Backup 0 0 v440_snap_oracle_new
12/17/2012 12:24 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
12/17/2012 12:23 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
12/17/2012 12:21 12/31/2012 5 2 N Full Backup 0 0 v440_snap_oracle_new
12/17/2012 11:22 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
12/17/2012 11:22 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
12/17/2012 11:20 12/31/2012 5 2 N Full Backup 0 0 v440_snap_oracle_new
12/17/2012 10:21 01/17/2013 1 8480 N User Backup 0 0 v440_snap_oracle_new
one more thing i used internal system hard-disk in place of SAN/NAS. vxvm&vxfs is configured on that machine.
hi Rleon,
when i run this script from netbackup oracle-client using ./jnbSA
creat the script "backup,archive And restore" option it runs successfull
script detail given below
INF - Begin progress logging for process: (4647)
INF - Using policy Oracle_v440 to perform a user directed Oracle Stream Based backup.
INF - Starting Oracle Recovery Manager.
INF - Using: /u01/oracle/102/bin/rman
INF - Connection info: target 'sys/*****@ORA10GR2' nocatalog
INF - Start of Recovery Manager input.
INF - # -----------------------------------------------------------------
INF - # RMAN command section
INF - # -----------------------------------------------------------------
INF - RUN {
INF - ALLOCATE CHANNEL ch00
INF - TYPE 'SBT_TAPE';
INF - SEND 'NB_ORA_CLIENT=v440,NB_ORA_SID=ORA10GR2,NB_ORA_SERV=Master-NTC';
INF - BACKUP
INF - INCREMENTAL LEVEL=0
INF - FORMAT 'bk_u%u_s%s_p%p_t%t'
INF - DATABASE;
INF - RELEASE CHANNEL ch00;
INF - # Backup Archived Logs
INF - sql 'alter system archive log current';
INF - ALLOCATE CHANNEL ch00
INF - TYPE 'SBT_TAPE';
INF - SEND 'NB_ORA_CLIENT=v440,NB_ORA_SID=ORA10GR2,NB_ORA_SERV=Master-NTC';
INF - BACKUP
INF - FORMAT 'arch_u%u_s%s_p%p_t%t'
INF - ARCHIVELOG
INF - ALL
INF - DELETE INPUT;
INF - RELEASE CHANNEL ch00;
INF - # Control file backup
INF - ALLOCATE CHANNEL ch00
INF - TYPE 'SBT_TAPE';
INF - SEND 'NB_ORA_CLIENT=v440,NB_ORA_SID=ORA10GR2,NB_ORA_SERV=Master-NTC';
INF - BACKUP
INF - FORMAT 'ctrl_u%u_s%s_p%p_t%t'
INF - CURRENT CONTROLFILE;
INF - RELEASE CHANNEL ch00;
INF - }
INF - End of Recovery Manager input.
INF - Start of Recovery Manager output.
INF - Begin progress logging for process: (4705)
INF - Start backup of backup piece: (bk_u9knt6kou_s308_p1_t802378526)
INF - Backup id: v440_1355820048
INF - Policy: Oracle_v440
INF - Master server: Master-NTC
INF - Beginning backup on server Media-NTC of client v440.
INF - End backup of backup piece: (bk_u9knt6kou_s308_p1_t802378526)
INF - Server status = 0
INF - Sun Microsystems Inc. SunOS 5.10 Generic January 2005
INF - You have new mail.
INF - Recovery Manager: Release 10.2.0.1.0 - Production on Tue Dec 18 18:55:22 2012
INF - connected to target database: ORA10GR2 (DBID=4104249781)
INF - using target database control file instead of recovery catalog
INF - allocated channel: ch00
INF - channel ch00: starting incremental level 0 datafile backupset
INF - input datafile fno=00001 name=/u01/oracle/oradata/ORA10GR2/system01.dbf
INF - input datafile fno=00003 name=/u01/oracle/oradata/ORA10GR2/sysaux01.dbf
INF - input datafile fno=00005 name=/u01/oracle/oradata/ORA10GR2/example01.dbf
INF - input datafile fno=00002 name=/u01/oracle/oradata/ORA10GR2/undotbs01.dbf
INF - input datafile fno=00004 name=/u01/oracle/oradata/ORA10GR2/users01.dbf
INF - piece handle=bk_u9knt6kou_s308_p1_t802378526 tag=TAG20121218T185525 comment=API Version 2.0,MMS Version 5.0.0.0
INF - channel ch00: backup set complete, elapsed time: 00:01:15
INF - Begin progress logging for process: (4705)
INF - Start backup of backup piece: (bk_u9lnt6kr9_s309_p1_t802378601)
INF - Backup id: v440_1355820124
INF - Policy: Oracle_v440
INF - Master server: Master-NTC
INF - Beginning backup on server Media-NTC of client v440.
INF - End backup of backup piece: (bk_u9lnt6kr9_s309_p1_t802378601)
INF - Server status = 0
INF - Begin progress logging for process: (4705)
INF - Start backup of backup piece: (arch_u9mnt6ks8_s310_p1_t802378632)
INF - Backup id: v440_1355820156
INF - Policy: Oracle_v440
INF - Master server: Master-NTC
INF - Beginning backup on server Media-NTC of client v440.
INF - End backup of backup piece: (arch_u9mnt6ks8_s310_p1_t802378632)
INF - Server status = 0
INF - channel ch00: starting incremental level 0 datafile backupset
INF - including current control file in backupset
INF - piece handle=bk_u9lnt6kr9_s309_p1_t802378601 tag=TAG20121218T185525 comment=API Version 2.0,MMS Version 5.0.0.0
INF - channel ch00: backup set complete, elapsed time: 00:00:26
INF - released channel: ch00
INF - sql statement: alter system archive log current
INF - allocated channel: ch00
INF - current log archived
INF - channel ch00: starting archive log backupset
INF - input archive log thread=1 sequence=8 recid=1 stamp=802378151
INF - input archive log thread=1 sequence=9 recid=2 stamp=802378631
INF - input archive log thread=1 sequence=10 recid=3 stamp=802378631
INF - piece handle=arch_u9mnt6ks8_s310_p1_t802378632 tag=TAG20121218T185712 comment=API Version 2.0,MMS Version 5.0.0.0
INF - channel ch00: backup set complete, elapsed time: 00:00:26
INF - channel ch00: deleting archive log(s)
INF - Begin progress logging for process: (4705)
INF - Start backup of backup piece: (ctrl_u9nnt6kt6_s311_p1_t802378662)
INF - Backup id: v440_1355820185
INF - Policy: Oracle_v440
INF - Master server: Master-NTC
INF - Beginning backup on server Media-NTC of client v440.
INF - End backup of backup piece: (ctrl_u9nnt6kt6_s311_p1_t802378662)
INF - Server status = 0
INF - archive log filename=/u01/oracle/flash_recovery_area/ORA10GR2/archivelog/2012_12_18/o1_mf_1_8_8f0x2nnz_.arc recid=1 stamp=802378151
INF - archive log filename=/u01/oracle/flash_recovery_area/ORA10GR2/archivelog/2012_12_18/o1_mf_1_9_8f0xkpvw_.arc recid=2 stamp=802378631
INF - archive log filename=/u01/oracle/flash_recovery_area/ORA10GR2/archivelog/2012_12_18/o1_mf_1_10_8f0xkqtm_.arc recid=3 stamp=802378631
INF - released channel: ch00
INF - allocated channel: ch00
INF - channel ch00: starting full datafile backupset
INF - including current control file in backupset
INF - piece handle=ctrl_u9nnt6kt6_s311_p1_t802378662 tag=TAG20121218T185742 comment=API Version 2.0,MMS Version 5.0.0.0
INF - channel ch00: backup set complete, elapsed time: 00:00:26
INF - released channel: ch00
INF - Recovery Manager complete.
INF - End of Recovery Manager output.
INF - End Oracle Recovery Manager.
now i used this script template in backup selection of policy it successful runing of sanpshot child thread. view details
12/18/2012 14:16:17 - Info dbclient (pid=5916) Initiating PFI Snapshot Rotation
12/18/2012 14:16:18 - Info dbclient (pid=5916) EXIT - PFI Snapshot Rotation command. Status = 0: the requested operation was successfully completed
12/18/2012 14:16:18 - Info dbclient (pid=5916) Initiating Create Snapshot using method auto.
12/18/2012 14:16:26 - Info dbclient (pid=5916) EXIT - Create Snapshot. Status = 0: the requested operation was successfully completed
12/18/2012 14:16:32 - Info dbclient (pid=5916) Initiating Delete Snapshot
12/18/2012 14:16:35 - Info dbclient (pid=5916) EXIT - Delete Snapshot. Status = 0: the requested operation was successfully completed
12/18/2012 14:16:35 - Info dbclient (pid=5916) the requested operation was successfully completed (0)
after that it next child process it fail error status 199 under job overview tab file list is empty and job details also empty.
don't know where m i doing wrong?
waiting ur kind reply
-rw------- 1 root root 8.3M Dec 18 12:54 v440_1355817277_C1_F1.1355817277.img
-rw------- 1 root root 1.0K Dec 18 12:54 v440_1355817277_C1_F1.1355817277.info
-rw------- 1 root root 1.0K Dec 18 12:55 v440_1355817308_C1_HDR.1355817308.img
-rw------- 1 root root 1.0K Dec 18 12:55 v440_1355817308_C1_HDR.1355817308.info
-rw------- 1 root root 8.3M Dec 18 12:55 v440_1355817308_C1_F1.1355817308.img
-rw------- 1 root root 1.0K Dec 18 12:55 v440_1355817308_C1_F1.1355817308.info
-rw------- 1 root root 1.0K Dec 18 13:33 v440_1355819576_C1_HDR.1355819576.img
-rw------- 1 root root 1.0K Dec 18 13:33 v440_1355819576_C1_HDR.1355819576.info
-rw------- 1 root root 361M Dec 18 13:33 v440_1355819576_C1_F1.1355819576.img
-rw------- 1 root root 1.0K Dec 18 13:33 v440_1355819576_C1_F1.1355819576.info
-rw------- 1 root root 1.0K Dec 18 13:33 v440_1355819616_C1_HDR.1355819616.img
-rw------- 1 root root 1.0K Dec 18 13:33 v440_1355819616_C1_HDR.1355819616.info
-rw------- 1 root root 254M Dec 18 13:34 v440_1355819616_C1_F1.1355819616.img
-rw------- 1 root root 1.0K Dec 18 13:34 v440_1355819616_C1_F1.1355819616.info
-rw------- 1 root root 1.0K Dec 18 13:34 v440_1355819646_C1_HDR.1355819646.img
-rw------- 1 root root 1.0K Dec 18 13:34 v440_1355819646_C1_HDR.1355819646.info
-rw------- 1 root root 8.3M Dec 18 13:34 v440_1355819646_C1_F1.1355819646.img
-rw------- 1 root root 1.0K Dec 18 13:34 v440_1355819646_C1_F1.1355819646.info
-rw------- 1 root root 1.0K Dec 18 13:40 v440_1355820048_C1_HDR.1355820048.img
-rw------- 1 root root 1.0K Dec 18 13:40 v440_1355820048_C1_HDR.1355820048.info
Did you create a User Backup schedule with appropriate backup window within your policy?
NetBackup status code: 199
Message: operation not allowed during this time period
Explanation: A user backup or archive was requested and this client is not in a
policy that has the following: a user backup or archive schedule with an open
backup window. This error implies that an appropriate policy and schedule
combination exists for this client.
Recommended Action: Determine the policies to which this client belongs that
also have a schedule of the appropriate type (either user backup or archive).
will restore -- where there is a Will there is a way
A few things.
I cannot see the word "proxy" anywhere in your RMAN script. This means that you are not doing a snapshot backup. Your backups are normal stream based backups.
Enabling the snapshot check box in the policy attribute would not enable snapshot backups unless the RMAN script is also done correctly. The messages you see about snapshots is probably about this issue. Notice how the snapshot was deleted within 9 seconds of its creation. That is way too fast for Nbu to have backed up any data from the snapshot.
Since you are trying to do a local snapshot backup on Solaris using VxFS, I am just guessing that you are trying to use the basic nbu_snap snapshot method. There are some prerequisites you will need to fulfill, such as configuring a cache device. Please refer to the Nbu Snapshot Client guide for details.
For other snapshot methods, and for setting up VxVM/VxFS based snapshot operations in general, you will have to refer to the Storage Foundation guides.
As wr pointed out, status code 199 is schedule related. Please check that your automatic and default application schedules are setup correctly according to the Nbu Oracle admin guide.
As for restoring snapshot based RMAN backups, the procedures are exactly the same as restoring normal stream based backups. So there is nothing to worry about with the restore procedures; if you are familiar with using RMAN to restore and recover.
The restore procedures are documented in both the Oracle and Nbu's guides. Basically, you have to create a RMAN restore script and run it. In fact you don't even need to touch Netbackup in order to run the restore script.
One of the best ways to troubleshoot Oracle RMAN related backup issues, is to just test the RMAN script without Netbackup. I.e., try your RMAN script with disk type channels so that it does not call Netbackup, and make sure that backups and restores are successful before getting Netbackup involved.
And again, unless you have one or both of the two reasons I listed in my first reply, there is usually very little reason to use snapshot based backups with Oracle.
thanx for reply Rleon & WU,
i configure this snapshot-policy via wizard and select the snapshot option. i share it's logs in a morning because here is 11:30 pm.
i will also try the methods that u advice me in ur reply.
As RLeon pointed out - the policy is one half of Oracle snapshot backups.
You REALLY need to go through all of the documentation that have been mentioned in RLeon's earlier posts. Trust me - it works when you follow steps in the manuals. Don't rely on wizards...
The 'Snapshot Client' section in NBU for Oracle contains many important configuration steps - like adding 'proxy' statement, the need for open backup window in BOTH Automatic and Application Backup schedules, as well as the need for stream-based backup for control files....
Supporting Storage Foundation and VCS on Unix and Windows as well as NetBackup on Unix and Windows
Handy NBU Links
hi,
its output of the restoration script.
INF - Begin progress logging for process: (26930)
INF - Putting the database in mount state.
INF - Using: /u01/oracle/102/bin/sqlplus
INF - Connection info: 'connect sys/*****@ORA10GR2 as SYSDBA'.
INF - Start of SQLPLUS output.
INF - Sun Microsystems Inc. SunOS 5.10 Generic January 2005
INF - You have new mail.
INF - SQL*Plus: Release 10.2.0.1.0 - Production on Wed Dec 19 10:58:32 2012
INF - Copyright (c) 1982, 2005, Oracle. All rights reserved.
INF - SQL> Connected.
INF - SQL> Database closed.
INF - Database dismounted.
INF - ORACLE instance shut down.
INF - SQL> ORACLE instance started.
INF - Database mounted.
INF - SQL> Disconnected from Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - 64bit Production
INF - End of SQLPLUS output.
INF - Starting Oracle Recovery Manager.
INF - Using: /u01/oracle/102/bin/rman
INF - Connection info: target 'sys/*****@ORA10GR2' nocatalog
INF - Start of Recovery Manager input.
INF - # -----------------------------------------------------------------
INF - # RMAN command section
INF - # -----------------------------------------------------------------
INF - RUN {
INF - ALLOCATE CHANNEL ch00
INF - TYPE 'SBT_TAPE';
INF - SEND 'NB_ORA_SERV=Master-NTC';
INF - RESTORE
INF - DATABASE;
INF - RECOVER
INF - DATABASE;
INF - RELEASE CHANNEL ch00;
INF - }
INF - End of Recovery Manager input.
INF - Start of Recovery Manager output.
INF - Begin progress logging for process: (27082)
INF - Start restore of backup piece: (bk_ubpnt8bvm_s377_p1_t802435062)
INF - Sun Microsystems Inc. SunOS 5.10 Generic January 2005
INF - You have new mail.
INF - Recovery Manager: Release 10.2.0.1.0 - Production on Wed Dec 19 10:59:13 2012
INF - connected to target database: ORA10GR2 (DBID=4104249781, not open)
INF - using target database control file instead of recovery catalog
INF - allocated channel: ch00
INF - Starting restore at 19-DEC-12
INF - channel ch00: starting proxy restore
INF - restoring datafile 00001 to /u01/oracle/oradata/ORA10GR2/system01.dbf
INF - proxy file handle=bk_ubpnt8bvm_s377_p1_t802435062
INF - restoring datafile 00002 to /u01/oracle/oradata/ORA10GR2/undotbs01.dbf
INF - proxy file handle=bk_ubpnt8bvm_s377_p4_t802435062
INF - restoring datafile 00003 to /u01/oracle/oradata/ORA10GR2/sysaux01.dbf
INF - proxy file handle=bk_ubpnt8bvm_s377_p2_t802435062
INF - restoring datafile 00004 to /u01/oracle/oradata/ORA10GR2/users01.dbf
INF - proxy file handle=bk_ubpnt8bvm_s377_p5_t802435062
INF - restoring datafile 00005 to /u01/oracle/oradata/ORA10GR2/example01.dbf
INF - proxy file handle=bk_ubpnt8bvm_s377_p3_t802435062
INF - channel ch00: proxy copy complete, elapsed time: 00:01:37
INF - Finished restore at 19-DEC-12
INF - Starting recover at 19-DEC-12
INF - starting media recovery
INF - media recovery complete, elapsed time: 00:00:04
INF - Finished recover at 19-DEC-12
INF - released channel: ch00
INF - Recovery Manager complete.
INF - End of Recovery Manager output.
INF - End Oracle Recovery Manager.
INF - Putting the database in open state.
INF - Using: /u01/oracle/102/bin/sqlplus
INF - Connection info: 'connect sys/*****@ORA10GR2 as SYSDBA'.
INF - Start of SQLPLUS output.
INF - Sun Microsystems Inc. SunOS 5.10 Generic January 2005
INF - You have new mail.
INF - SQL*Plus: Release 10.2.0.1.0 - Production on Wed Dec 19 11:01:03 2012
INF - Copyright (c) 1982, 2005, Oracle. All rights reserved.
INF - SQL> Connected.
INF - SQL>
INF - Database altered.
INF - SQL> Disconnected from Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - 64bit Production
INF - End of SQLPLUS output.
i already tested the vxfs using nbu_snap option its work and now i am going to test its backup for restoration on different machine.
once done or i face any issue then upload its logs accordingly.
Seems your question in your opening post was answered, right?
Please mark one of RLeons posts as solution.
When you test restore on different machine and instructions in the manual are not helping, please start a new discussion for the new issue.
Supporting Storage Foundation and VCS on Unix and Windows as well as NetBackup on Unix and Windows
Handy NBU Links
Seems your question in your opening post was answered, right?
yup
----------
thanx all
if i face further problem i will open new discussion.
Would you like to reply?
Login or Register to post your comment.