HP Data Protector Operations Guide About this document ..........................................................................................................................3 Data Protector architectural overview ...................................................................................................3 Data Protector Cell .........................................................................................................................3 Cell Manager services ............................
DCBF ......................................................................................................................................37 Tablespaces.............................................................................................................................39 IDB notifications and reporting ...................................................................................................41 Recovering the IDB ..............................................................................
About this document This document is intended for backup, system or storage operators and administrators, who are new to Data Protector and are performing common backup tasks. It covers frequently performed maintenance tasks, and provides some configuration recommendations and best practices on how to set up an effective and efficient backup environment. This is not intended to replace any existing documentation. For other Data Protector documentation, please refer to www.hp.com/support/manuals.
The Data Protector client/server architecture provides multiple manager layers, which offer tremendous flexibility and adjust easily to organizational needs and changes. Cell Manager and clients The Cell Manager is the heart of the Data Protector backup environment. The clients are controlled from the Cell Manager system. Enterprise Console The Data Protector integration with HP Operations Manager provides the concept of the Enterprise Console.
Notes: • See the HP Data Protector concepts guide (B6960-90151) for further information on cell architecture. • See the Cell Manager Planning and Sizing Guide (4AA2-5036ENW) and Capacity Planning Spreadsheet for further details on cell sizing.
Verify the services status by running the following command line option: The Session Manager The Cell Manager listens for session requests and starts the appropriate Session Manager, which in turn starts the required clients. A dedicated Session Manager controls the clients for each operation. If a new session is started, an additional Session Manager is generated.
These 3 default accounts can be removed if required, as long as another account has been defined with access to all Data Protector clients and user rights.
The User has the following default user rights: New user groups can be created with custom user rights. Using a service account The CRS service on windows has an owner assigned with certain permissions and a password, which needs to be updated if the user password changes. Use a dedicated service account if you do not want to change the passwords. Wild-card user Caution: For security reasons, it is not recommended to add a wild-card user. It is only recommended for use in test environments.
To add the any user, click on Add User, then fill in Any under Name, Domain and/or Client: About user management Data Protector users are based on the operating system user. Data Protector backup session ownership is based on a session level. This means that if there are multiple clients in a session or backup spec, it is not possible to split the ownership. Backup specifications need to be organized so there are never two clients with different owners in the same backup spec.
Note: It is probably easier to directly modify the C:\ProgramData\OmniBack\Config\Server\users\userlist file to add batches of users than to do it through the GUI.
RMAN, in conjunction with the Data Protector oracle integration, can perform backups and restores. In order for RMAN initiated backups and restores to be successful, you need to add certain other users to Data Protector: • • • For Data Protector 6.1 on UNIX systems, add the users root and the oradba account. For Oracle, you need to add the RAC root and oracle dba account. For Data Protector 6.11 on UNIX systems, you only need to add the oradba account.
Data Protector media management automatically selects the most appropriate media for backup. Basic media selection criteria: • If available, media in good condition are used first. • Media in fair condition are used only if no media in good condition are available. • Media in poor condition are not selected for backup. • Media are always selected from the specified pool. If the pool does not contain unprotected media, Data Protector accesses a free pool (if configured).
Resolving mount requests Data Protector issues a mount request if either it requires a specific medium to read data from, or it needs more media but none are available in the device. To resolve the mount request, add additional media or cancel the device: • To confirm the mount request, insert the required medium into the device and click Confirm Mount Request. Alternatively, use omnimnt on the Data Protector CLI to confirm the mount request.
The following screenshot shows the media that is added to the pre-allocation list: Using a free pool A free pool is a media pool that you can configure to allow free media to be shared across media pools, which may reduce operator intervention due to mount requests. The use of a free pool is optional. A free pool: • cannot be deleted if it is linked with a media pool or if it is not empty. • is different from a regular pool as it cannot be used for allocation because it cannot hold protected media.
• • You may experience some temporary inconsistencies (1 day) in pools when using free pools (for example when there is an unprotected medium in a regular pool waiting for deallocation to the free pool). If a free pool contains media with different data format types, Data Protector automatically reformats allocated media if necessary. For example, NDMP media may be reformatted to normal media. To create a free pool, right-click on Media, Pools and select Free Pool.
| P a g e
Note: Invalid media will be removed by Data Protector: Backup management This chapter covers hints and tips about common backup management tasks, for instance creating and viewing a backup specification, and soon. It explains a few HP Data Protector internals, such as using drive concurrency and multiplexing. It describes how to configure and run reports, how to monitor backup sessions, and how to resolve failed sesisons.
Click on the View menu to select what criteria to use for viewing the backup specifications: Right-click to add a group, and specify the name of the new group: Click Change-Group… to move backup specifications into their dedicated groups: Adding a new backup specification An existing backup specification can be quickly copied and edited through the Data Protector GUI.
To copy a backup specification, right-click and chose Copy As…: Using drive concurrency The number of Disk Agents started for each Media Agent is called Disk Agent (backup) concurrency and can be modified using the Advanced options for the device or when configuring a backup. Note: The concurrency set in the backup specification takes precedence over the concurrency set in the device definition. Data Protector provides a default number of Disk Agents that are sufficient for most cases.
Order of Disk Agents started Data Protector has 2 backup modes: SAN Backup mode, and LAN Backup mode. 1. If you do a SAN backup, Data Protector always tries to run local backups. It will fill up all slots of a running local Media Agents with available Disk Agent slots.
Defining backup specification-based drive concurrency The backup object summary shows the number of Disk Agents configured for a filesystem backup specification. To add a new Disk Agent, go through the Manual add… option, and follow the wizard to specify the filesystem backup details: Chose which type of object is used for the backup specification. Then select the client and mount point, optional filters and reporting parameters, advanced properties and other filesystem options.
Note: You can also specify a new raw disk partition via this option in the GUI. Click on Manual Add, and chose a Disk Image Object. About multiplexing Multiplexed media contain interleaved data of multiple objects. Such media may arise from backup sessions with a device concurrency of more than 1. Multiplexed media may compromise the privacy of backups and require more time for restore. Using the Data Protector object copy functionality, you can demultiplex media.
During the copy operation definition, a number of parameters can be customized for backupand catalog protection, recycling source copies, and ejecting media after a successful copy: Schedule the copy session, or launch an interactive copy. The post-backup copy session will run after the backup session has been completed. For an interactive copy, chose the session that needs to be copied: Select the source session version from the interactive copy wizard.
After the session has run successfully, the session messages in the IDB will show a (copy) backup type: The backup type will show as (copy) type for an object copy session. Object copy The Data Protector object copy functionality enables you to copy selected object versions to a specific media set. You can select object versions from one or several backup sessions or object consolidation sessions.
Object copy sessions can be run interactively, or scheduled, based on media, sessions, or objects. Select what type of copy operations you want to perform. Note: See the HP Data Protector concepts guide (B6960-90151) for further information on object copy. Copy session start time The copy session start time will always be the time of the original backup session.
It is not possible to calculate the exact duration of an object copy on a per object level. Emailing backup session reports A schedule can be added at the report group level. If only 1 report needs to be scheduled to be sent, create a different group with a single report for each schedule. Setting up webbased reporting Data Protector’s web-based reporting allows you to view reports on backup, object copy, and object consolidation status and Data Protector configuration, using the web-interface.
• • UNIX: /opt/omni/java/bin Windows: C:\Program Files\Omniback\java\bin In a browser on any system with access to the web server, open the following file from the copied java folder on the web server to display the Data Protector reporting: • UNIX: /bin/webreporting.html • Windows: C:\Program Files\Omniback\bin\WebReporting.html Make this file available to the users of the web reporting in the full URL form. For example, put a link to this file from your Intranet site.
To add media to the location, right-click on the media properties, and select Change Location. Hold down the Shift key to change the location for several media at once. The location will show the number of media added. Each medium will also show the location details. Performance monitoring using a nul device Backup performance numbers are displayed at the end of each backup session as a summary. Data Protector does not offer interactive performance monitoring for ongoing sessions through the GUI.
3. Create a backup specification. In the Options wizard page, set data protection to None and catalog protection to Same as data protection. Select the option Display Statistical Info to see the performance summary at the end of the backup session. 4. Perform backups to this nul device and check if the performance discrepancy between backups to the file device and backups to the real device can be explained.
Restarting failed sessions You can restart a failed session or a completed session with failures after you have resolved related problems. This restarts only the failed objects. The option can be used for clustered failed objects as well. You cannot restart failed sessions that are the result of an unsaved backup specification. To restart sessions: 1. In the Context List, click Internal Database. 2. In the Scoping Pane, expand the Internal Database item and click Sessions. 3.
The following screenshot shows how to resume a failed session. Right-click on the session name, and chose Resume Session: Editing the backup schedules You cannot edit a backup schedule created in Data Protector from the GUI. You need to delete and recreate it. You can edit it via the schedule template file by following the required format. Bulk edits are also quicker done from the templates than the GUI.
Note: Use omnitrig –stop to stop the scheduler: Restore management This chapter covers file version and file search based restore operations, and performing a restore after a library configuration has been deleted. File version restore File version restore is available via the backup session history, or via the backup objects. The backup objects will show the backup sessions for the file: full, incremental, enhanced incremental, or synthetic or virtual full session details.
Restore query To search for a file and version backed up, go to the Task list under the Restore options, and use the wizard to find the file.
• For File system backups only, you can achieve the same by specifying the –device option to omnir on the command line. For integrations backup, you can use other alternatives. Keeping a restoredev file under the directory: o UNIX: /etc/opt/omni/server/cell o Windows: \Config\Server\cell This is a plain text file containing the old and new device names separated by a space. It is referred to whenever the device is called, and the old name is replaced by the new name.
Every day at 12:30 P.M. by default, Data Protector starts checks for the following: • IDB Space Low • IDB Tablespace Space Low • Not Enough Free Media • Health Check Failed • User Check Failed (if configured) • Unexpected Events • License Warning • License Will Expire • IDB Purge Needed By default, any triggered notification is sent to the Data Protector Event Log.
Uncomment the option in the global file, and select 0 or 1 to enable the option: Instead of the automatic scheduled purge preview, it is recommended to script or manually run a purge preview session once a month, using the following command: C:\Program Files\OmniBack\bin>omnirpt –report db_purge_preview Analyze the output of the db_purge_preview report by looking at the column Est. Obs.
Example: Note: C:\Program Files\OmniBack\bin>omnidbutil –purge_stop This is useful command to stop a purge in case you need to run urgent Data Protector operations.
Run the following command to determine the layout of the dcbf directories and files: • UNIX: /opt/omni/sbin/omnidbutil –list_dcdirs • Windows: C:\Program Files\OmniBack\bin>omnidbutil -list_dcdirs Analyze the output and modify all existing dcbf directories if they do not meet the above recommendations, using the following command: • UNIX: /opt/omni/sbin/omnidbutil (–modify_dcdir • /var/opt/omni/server/db40/dcbf (–maxsize 32768 (–maxfiles 10000 (– spacelow 100 Windows: C:\Program Files\OmniBack\bin>omnid
Tablespaces It is recommended to check the size of the following tablespaces: fnames.dat, fn1.ext, fn2.ext, fn3.ext, fn4.ext and dirs.
Each tablespace listed must have enough space otherwise you will encounter the no Log message and will not be able to select individual files for restore. Analyze the output to see if any of the tablespaces need to be extended by comparing the Maximum size and Current size for each tablespace. If they are within 500 MB of each other, extend the tablespace using the following example command: • UNIX: /opt/omni/sbin/omnidbutil –extendtblspace fn2.
IDB notifications and reporting Setup report notification for IDB maintenance related tasks. Recovering the IDB Several recovery methods are available for recovering the Internal Database. Depending on the identified level of corruption, your requirements, and the availability of the IDB recovery file and the original device and transaction logs, the recovery procedure can differ.
Restore: Current situation Remark Recovery procedure (restoring IDB) The IDB recovery file is available but the original device used for the IDB backup has changed. The method is essentially the same as the guided autorecovery method, but less guided, more complex, and time consuming. Restore the IDB Using IDB Recovery File and Changed Device The IDB recovery file is not available. The method is essentially the same as the guided autorecovery method, but less guided, more complex, and time consuming.
mv /var/opt/omni/server/db40 /var/opt/omni/server/db40.save mv /db40 /var/opt/omni/server/db40 7. The restore process also restored the configuration files into the same location as the database files. You may want to move them into place as well if they need to be recovered. Note: this step may be optional, if the files are intact. For Windows Cell Managers, use the windows explorer. For Unix use the following commands: mv /etc/opt/omni/server /etc/opt/omni/server/omni.
Other frequently used maintenance related global options are the following: • dcbf related options 9 DCDirAllocation=0, 1, 2 9 MaxDCDirs=NumberOfDirectories 9 SessionMessagesDir=FullPathToTheMessageDir • cdb related options 9 9 DBFreeDiskSpace=MinSpaceInMBytes DBFreeExtFileSpace=MinSpaceInMBytes • general options 9 RecoveryIndexDir=FullPathToTheBackupDir 9 DbXXXXXXXXXLimit=GBytes 9 DBPurgeSuspension=0 or 1 9 DBPurgeSuspensionDuringDBCheck=0 or 1 9 DailyMaintenanceTime=HH:MM 9 DailyCheckTime=HH:MM omnirc
scsitab file It is recommended that you let Data Protector configure backup devices automatically. Data Protector can automatically configure most common backup devices, including libraries. You still need to prepare the media for a backup session, but Data Protector determines the name, policy, media type, media policy, and the device file or SCSI address of the device, and also configures the drive and slots. You can also configure a backup device manually.
The cell_info file can be found in the following directory: C:\Program Files\OmniBack\Config\Server\cell Important: cell_info is a file created and edited during installation. Do not manually edit the file. Variables currently undocumented Treewalk A treewalk is performed when backing up a file system to calculate how many files have changed since the last full backup. NOTREEWALK=1 is the correct variable for Data Protector 6.0. It works exactly the same in the Data Protector 6.
7. 8. Right-click SNMP Service and select Properties. a. Select the Traps tab. Enter public in the Community name text box and the hostname of the Management Server in the Trap Destinations text box. b. Select the Security tab. Under Accepted community names, select the community public, click Edit and set Community rights to READ CREATE. c. Confirm your settings. Run omnisnmp. Note: The community name is case sensitive. To finish the setup, there are a few additional configuration steps.
Example before editing the OVdests file: Example after editing the OVdests file: Adding a community name registry key other than public Executing the omnisnmp command will create the Data Protector registry keys required. Action: An optional additional registry key entry can be created to add a community name other than public. If it is NULL, public is assumed as a value for the registry key. If traps need to be sent to the public community name, no entry is necessary.
In the example below, the local Cell Manager hostname or Cell Manager IP address is added, as well as the hostname or IP address of the remote trap receiver destination host. The Cell Manger name/address in the security tab is necessary. If you set it to ‘Accept SNMP traps from any host’, then no entries are necessary. For further information, refer to SNMP Configuration on Windows in chapter 2 of the HP Data Protector A06.10 integration guide for HP Operations Manager for Windows.
4. Configure Data Protector to send SNMP traps to the Operations Manager Server system: a. Using the Data Protector GUI Reporting context, set up all notification events to use: • SNMP as delivery method • Operations Manager Server system as the destination. b. Add the Operations Manager Server hostname as trap destination to the OVdests file in Data Protector Root/Config/server/SNMP. c. Disable filtering of SNMP traps by emptying the OVfilter file in Data Protector Root/Config/server/SNMP.
To verify what files have been backed up before a session aborted or failed, specify the session name with the session ID and the –report option: To view the session catalog information, specify the session ID and the –catalog option: To find which backed-up objects are available: C:\Program Files\OmniBack\bin>omnidb –object To find the backup sessions in the database: C:\Program Files\OmniBack\bin>omnidb –session 51 | P a g e
To perform a query of a specific session: C:\Program Files\OmniBack\bin>omnidb –session 2009/09/10-18 To look at a detailed session report: C:\Program Files\OmniBack\bin>omnidb –session 2009/09/10-18 –detail To see a list of files backed up during a specific session: C:\Program Files\OmniBack\bin>omnidb –winfs “haptic.xst.rose.hp.
If the object is using a backup description, this needs to be specified on the command line as well.
devbra To verify what devices are visible to the host, use the following command line option: Note: As an alternative is the devbra command, you can use the HP StorageWorks Library and Tape Tools (HP L&TT): http://h18006.www1.hp.com/products/storageworks/ltt/index.html Log files and troubleshooting Data required for support calls The session log The Data Protector session log lists error messages. Click on an error to get more details.
To look at the failed session messages after the session window has been closed, go to the Internal Database view, and find the session message window.
Support files The table below describes the Data Protector log files: Log File debug.log inet.log enhincr.log Ob2EventLog.txt media.log omnisv.log security.log purge.log RDS.log Description Contains unexpected conditions. While some can help you, the information is mainly used by the support organization. Contains local security-related events for the client, such as denied requests. On UNIX, it also contains all requests made to the Data Protector Inet service.
Debugging Data Protector Almost all Data Protector commands can be started with an additional -debug parameter that has the following syntax: –debug 1–99[,C:][,T:][,U] [] Where: • 1–200 is the debug range. Specify the range 1–200 unless instructed otherwise. Specify optional parameters as a part of the range parameter, separated by commas: o C: limits the size of debug files to n kilobytes. The minimum value is 4 (4 kB) and the default value is 1024 (1 MB).
Click on Use these settings for the next restart only, and click Restart now… The Data Protector debug log files will be located under: • • Unix: /tmp Windows: C:\Program Files\Omniback\tmp You can change the location with omnirc option OB2DBGDIR: # # # # # # # # OB2DBGDIR= Default: none This variable is used to change the location of debug files on a per system basis. You have to specify a fully qualified path of an existing directory.
Use the command line debug log collector to zip up the debug files from Cell Manager and clients. To unpack debug files that have been zipped on a UNIX Cell Manager, on a Windows system, copy the omnidlc.exe file over to a Windows system, and unpack the *.pck files, running omnidlc –unpack. Note: This is an undocumented and unsupported operation.
Example: C:\Program Files\OmniBack\bin>telnet caspase 5555 HP Data Protector A.06.11: INET, internal build 243, built on Tuesday, August 25, 2009, 7:08 AM Patch upgrade and versioning Patches can be pushed from the Cell Manager or Installation Server GUI, or installed locally from the CDs or DVD. Right-click on the host name to chose Add Components or Upgrade, and select the components that need to be installed.
Clients will verify the source for each request and allow only those requests received from clients selected in the Enable Security on selected client(s) window. These clients are listed in the allow_hosts file. If the request is denied, the event is logged to the inet.
Operation audit checklist Backing-up data Control Objective Backup concepts Backup operations Procedure Result Before you backup, review key concepts and requirements. Determine where you will store the backup. Determine which files, folders, or volumes you want to back up and whether the backups will need to be used for operating system (critical volumes only), full server (all volumes), system state, or bare metal recovery. Determine how many times a day and at what times you want to run backups.
rotational procedure by identifying the type and level of backup, which generation is moved offsite, how many generations are retained off-site, and which day the rotation occurs. For each environment, obtain screen captures from the backup software that show: • The selection of files that are backed up. • The schedule that the backup job is set to follow. • A recent log file showing a successful backup of the system. • A recent restore log file (if available) showing a successful restore of the system.
Short-term maintenance checklist Control Objective Backup maintenance Media maintenance Procedure Result Check the Data Protector Event log for daily notifications. Restart failed backup sessions. Resume failed sessions. Verify media and pool usage. Resolve poor media issues. Long term maintenance checklist Control Objective Database maintenance Log files Procedure Result Run the IDB purge operation. Analyze DCBF directories’ capacity usage. Check the size of the tablespaces.
misappropriation of the information stored, and is reasonably accessible during non-business hours. References www.hp.
© Copyright 2008 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein. Linux is a U.S. registered trademark of Linus Torvalds.
