Loading...

Advanced Backup - Documentum iDataAgent (Oracle)

Table of Contents

Overview

What Happens during Documentum Backups

Incremental Backups

Database Backups

Storage Area Backups

Full-Text Index Backups

Managing Jobs

Additional Options

Overview

Once you have performed a full backup of the Documentum Repository initially, the subsequent backups can be either Full backups or Incremental backups.

A full backup contains all the data that comprises a subclient. Though, full backups provide the most comprehensive protection of data, it consumes the most amount of time and resources. A full backup is supported for backing up database files, Storage Areas, and Full-text Indexes. Optionally, log files can be backed up as well.

To streamline the backup process and to optimize the time and resource, you can run incremental backups. You can perform these backups from the CommCell Console. Incremental backups are supported for backing up database files and Storage Areas.

Backups can be run on demand or can be scheduled to be run at a later point of time. However, it is always recommended that you schedule your backups at frequent intervals. For example, full backups can be scheduled on a weekly basis and incremental backups can be scheduled on a daily basis.

The following sections describes the various types of backups supported by this agent.

What Happens during Documentum Backups

In order to achieve a consistent backup of Documentum components, Commvault Systems, Inc. performs the following during a repository backup:

  1. Pre-backup phase includes:
    1. Quiesces the Full-Text index (indexer process).
    2. Inactivate dm_DMclean and dm_DMfilescan jobs.
    3. Create Commvault Systems, Inc. Index.
  2. Backs up the Full-Text Index (FTI).
  3. Backs up the database.
  4. Backs up the log files.
  5. Backs up the storage area.
  6. Post-backup phase includes:
    1. Unquiesces the Full-Text Index (indexer process).
    2. Activates  dm_DMclean and dm_DMfilescan jobs.
  7. Archives Commvault Systems, Inc. Index and copies it to media.

Incremental Backups

An incremental backup contains only data that is new or has changed since the last backup, regardless of the type. On average, incremental backups consume far less media and place less of a burden on resources than full backups.

Always ensure that your first backup is a full backup that will contain all the data that comprises a subclient's contents.

The illustration on the right clarifies the nature of full and incremental backups. For simplicity, assume there is a file system that contains six files as represented in the figure.

Backup #1 is a full backup and therefore writes all the data, changed and unchanged, to the backup media. Backups #2 through #n-1 are incrementals and only back up those files that have changed since the time of the last backup, regardless of the type. For example, files A, B, and E changed after the full backup and were therefore backed up in Backup #2. Backup #4 backed up files A and D because both files were modified sometime after Backup #3 occurred. File F did not change; hence, it was not backed up in any of the incremental backups, but it was included in both full backups, which, by definition, back up everything.

By default, during the database backup phase of an incremental backup, only logs are backed up. This can be overridden by setting the sINCREMENTALDCTMDBBACKUP additional setting to Y.

When a Storage Area is backed up without any new documents, the backup job completes with the error code shown below but nothing is backed up as part of an incremental backup.

Error code: No new documents have been added since the last FULL Backup.

Starting an Incremental Backup

  1. From the CommCell Browser:
    •  Navigate to Client Computers | <Client> | Documentum | <Instance>.
    • Right-click the default subclient and click Backup.
  2. By default, Immediate is selected as backup type.
  3. In the Job Initiation section select the Backup Schedule.
    • Click Immediate if you want to run the backup now.
    • Click Schedule if you want to schedule the backup for a specific time.
  4. Click OK. You can track the progress of the job from the Job Controller window of the CommCell console. Once the job is complete, view the job details from the Backup History. Right-click the Subclient and select Backup History.

Database Backups

The Documentum iDataAgent enables you to backup the database in different environments based on need. Database files can be backed up in full or incremental backups. These are described in the following sections.

Notes:

  • If the database and storage areas are backed up with separate subclients, ensure to schedule the backups so that the database backup is performed first followed by the storage area and full-text index backups.
  • Use workflow to back up the SQL databases.

Offline Backups

Offline backups are performed when the database is shut down and unavailable for use. Offline backups are always full backups and do not include the archived log files.

In order to backup the database when it is offline, you can create a separate user-defined subclient for offline backup. See Creating Subclient to Backup Offline Databases for step-by-step instructions.

Note: When performing Documentum offline backups from the CommCell Console, shutdown the Documentum Repository services before performing the backup. This can also be run as a Pre Process during backing up to shut it down gracefully. A post backup script can then be run to start it up again after running the backup.

Online Backups

Online backups are performed when the database is online. Oracle databases must be in the ARCHIVELOG mode. Data files are included when you back up the database in online mode. You can perform a full or incremental backup when the database is online. These backups do not require database downtime and are extremely useful when you need to perform a point-in-time restore of the database.

In order to backup the database when it is online, you can create a separate user-defined subclient for online backup. See Creating Subclient to Backup Online Databases for step-by-step instructions.

Archive Log Backups

Archive log backups are useful when you want to recover database transactions that have been lost due to an operating system or disk failure. You can apply these archive logs to an online backup in order to recover a database.

Full backups generally includes both the data and logs. However, you can also take backups of the archive logs only. In order to backup only the archive logs, you need to create a separate subclient for archive logs. See Creating Subclient to Backup Archived Log Files for step-by-step instructions.

By default, the first backup operation on the default subclient will include the entire database (data and archive logs) when the database is online and in ARCHIVELOG mode for Oracle databases.

Frequent backups of database log files should be performed. Without them, a database can be recovered only to the time of its last full backup.

Control File and SP File Backups

Including Control Files in a backup is useful since the control file stores the status of the physical structure of the database. It is absolutely crucial to database operation. Control files keep track of data files, control files should be backed up whenever a change is made to the database.

Server Parameter (SP) files control database resources and is basically a repository for initialization parameters. It is maintained on the client computer where the Oracle database resides and is read at system startup. It should be included in a backup, especially if Recovery Manager (RMAN) is not used to back it up in the backup strategy of an Oracle database.

Selective Copy - Full Backups

A selective copy allows you to copy backup data selectively from a source copy to this copy, providing for better tape rotation. Since only selective backups can be copied to selective copies, the selective copies cannot be promoted to the primary copy, only synchronous copies can be promoted. Note that the data selection process does not have to be the same for all auxiliary copies.

During an auxiliary copy operation or a data protection operation that creates an inline copy, only those backups from the primary copy that meet certain criteria will be copied to a selective copy. You can define a selective copy to be time-based, automatically selected, or automatically not selected on the primary copy.

If the copy is defined as All Fulls, all full backups on the primary copy will be copied during an auxiliary copy operation or a data protection operation that creates an inline copy. If the copy is defined as time-based, only the first or last full backup that occurs within each selected weekly, monthly, quarterly, half-yearly, or yearly interval will be copied. You have the option of creating and associating a custom calendar to the copy, so that the intervals can be further customized.

Selective copy will not only copy Selective Online Full and Selective Offline Full jobs, it will also copy regular Full jobs with the linked Log backup jobs.

You need to create a separate subclient to perform a selective copy full backup of the database. See Configuring Subclient for Selective Copy with Full Backups for step-by-step instructions.

Selective Copy is applicable only for jobs that have database backed up as part of full backup operations.

Storage Area Backups

The Documentum iDataAgent enables you to back up both local and distributed Storage Areas that are  distributed across different computers. With distributed storage areas, the File System iDataAgent should be installed on the client computers that are hosting the Storage Areas.

Both full and incremental backup types are supported for Storage Areas. The first backup of a Storage Area is performed as a Full backup. Storage Areas can be backed up in a full Repository backup or as a separate backup.

A pre or post backup process can be run when backing up Storage Areas.

Note: In order to discover far stores and their associated clients in a distributed environment, ensure that they are configured correctly.  

Full-Text Index Backups

The Documentum iDataAgent enables you to back up Full-text Indexes. The File System iDataAgent should be installed on the client computers that are hosting the Full-text Indexes.

Full-text Indexes can be backed up only in full backups. Even if an incremental backup job is run, Full-text Indexes will always be backed up as a full backup. Full-text Indexes can be backed up in a full Repository backup or as a separate backup. For the latter case, see Create Subclient to Only Contain Full-text Indexes for step-by-step instructions.

When backing up a Full-Text Index (FTI), it necessary to quiesce the Index Server to ensure consistent backups. For this purpose, quiesce and unquiesce scripts are required. These are used to quiesce and unquiesce the Indexer process in the Index Server during the backup operation respectively.

These two sample scripts are provided in the <installation_folder>/Base folder on the computer where the Documentum iDataAgent has been installed:

cvfti_shutdown.sh

cvfti_startup.sh

These scripts have to be modified to suit the environment prior to back up.

You can run a Full Text Index backup, using either FTI or xPlore as the indexing server.

Documentum FTI as the Indexing Server

If the Documentum indexing server is FTI, use the sample scripts given below. You can find these sample scripts in the <installation_folder>/Base folder on the computer where Documentum Agent has been installed.

To quiesce the Documentum Indexing Server

cvfti_shutdown.cmd

To unquiesce the Documentum Indexing Server

cvfti_startup.cmd

For Windows based Documentum Agents, the FTI quiesce and unquiesce scripts are similar to the scripts in the following example. These scripts have to be modified prior to backup to suit your environment:

cvftishutdown.bat

E:\Documentum\fulltext\IndexServer\bin\nctrl.exe stop

cvftistartup.bat

C:\WINDOWS\system32\net.exe start FASTDSService

E:\Documentum\fulltext\IndexServer\bin\nctrl.exe start

Documentum xPlore as the Indexing Server for UNIX

For UNIX based Documentum Agents, the FTI quiesce and unquiesce scripts are similar to the scripts in the following example. These scripts have to be modified prior to backup to suit your environment:

To quiesce the Documentum Indexing Server for UNIX

cvxplore_suspend.sh

#!/bin/sh

#Sample script for suspend Full Text Index server for warm backup

#It suspsends the index server

#This script has to be modified to suit the user environment.

export CV_XPLORE_LOG=/tmp/CV_xPlore.log

export errorlevel=0

echo "CALL Index Agent quiesce"

export DSEARCH_HOME=/home/docadmin/xPlore

export today=`date +%Y-%m-%d.%H:%M:%S`

echo "Index Agent quiesce begin at $today...." >> $CV_XPLORE_LOG

$DSEARCH_HOME/dsearch/admin/xplore.sh "suspendDiskWrites()"

export PH_1=$errorlevel

if [ "$PH_1" -eq 0 ] ; then

echo Index Agent quiesce is sucess $today... >> $CV_XPLORE_LOG

echo "Index Agent quiesce or DB suspend passed." >> $CV_XPLORE_LOG

else

echo "Index Agent quiesce is failed $today...." >> $CV_XPLORE_LOG

echo "Index Agent quiesce or DB suspend failed." >> $CV_XPLORE_LOG

fi

exit $errorlevel

To unquiesce the Documentum Indexing Server for UNIX

cvxplore_resume.sh

#!/bin/sh

#Sample script for resuming Full Text Index server (xplore).

#Warm backup requires to keep index server in suspend mode and resume

#after backup operation.

#This script has to be modified and make sure suitable for the environment.

export CV_XPLORE_LOG=/tmp/CV_xPlore.log

export errorlevel=0

echo "CALL Index Agent Unquiesce"

export DSEARCH_HOME=/home/docadmin/xPlore

export today=`date +%Y-%m-%d.%H:%M:%S`

echo "Index Agent Unquiesce begin at $today...." >> $CV_XPLORE_LOG

$DSEARCH_HOME/dsearch/admin/xplore.sh "resumeDiskWrites()"

export PH_1=$errorlevel

if [ "$PH_1" -eq 0 ] ; then

echo Index Agent Unquiesce is sucess $today... >> $CV_XPLORE_LOG

echo "Index Agent Unquiesce or DB Resume passed." >> $CV_XPLORE_LOG

else

echo "Index Agent Unquiesce is failed $today...." >> $CV_XPLORE_LOG

echo "Index Agent Unquiesce or DB Resume failed." >> $CV_XPLORE_LOG

fi

exit $errorlevel

Documentum xPlore as the Indexing Server for Windows

For Windows based Documentum Agents, the xPlore quiesce and unquiesce scripts are similar to the scripts in the following example. These scripts have to be modified prior to backup to suit your environment:

To quiesce the Documentum Indexing Server for Windows

cvxploreshutdown.bat

@REM set GALAXY_HOME=C:\TEMP

@REM set GALAXY_HOME=C:\TEMP

@REM set GALAXY_LOG=%GALAXY_HOME%

set CV_XPLORE_LOG=C:\TEMP\CV_xPlore.log

set errorlevel=0

@REM CALL Index Agent Unquiesce

echo "Index Agent quiesce begin at %date% %time%...." >> %CV_XPLORE_LOG%

@@ "C:\xPlore\dsearch\admin\xplore.bat suspendDiskWrites()"

set PH_1=%errorlevel%

IF NOT %PH_1% EQU 0 (

echo Index Agent quiesce Failed reason >> %CV_XPLORE_LOG%

GOTO FAIL

) ELSE (

echo "Index Agent quiesce is done at %date% %time%...." >> %CV_XPLORE_LOG%

)

:PASS

echo Index Agent quiesce or DB suspend Passed.

exit /b %errorlevel%

:FAIL

echo Eecho Index Agent quiesce or DB suspend Failed.

exit /b %errorlevel%

To unquiesce the Documentum Indexing Server for Windows

cvxplorestartup.bat

@REM set GALAXY_HOME=C:\TEMP

@REM set GALAXY_HOME=C:\TEMP

@REM set GALAXY_LOG=%GALAXY_HOME%

set CV_XPLORE_LOG=C:\TEMP\CV_xPlore.log

set errorlevel=0

@REM CALL Index Agent Unquiesce

echo "Index Agent Unquiesce begin at %date% %time%...." >> %CV_XPLORE_LOG%

@@ "C:\xPlore\dsearch\admin\xplore.bat resumeDiskWrites()"

set PH_1=%errorlevel%

IF NOT %PH_1% EQU 0 (

echo Index Agent Unquiesce Failed reason >> %CV_XPLORE_LOG%

GOTO FAIL

) ELSE (

echo "Index Agent Unquiesce is done at %date% %time%...." >> %CV_XPLORE_LOG%

)

:PASS

echo Index Agent Unquiesce or DB Resume Passed.

exit /b %errorlevel%

:FAIL

echo Either Index Agent Unquiesce or DB Resume Failed.

exit /b %errorlevel%

Managing Jobs

Once you initiate the backup operation, a backup job is generated in the Job Controller. Jobs can be managed in a number of ways. The following sections provide information on the different job management options available:

Restarting Jobs

Jobs that fail to complete successfully are automatically restarted based on the job restartability configuration set in the Control Panel. This configuration can be changed at any time; however, changes to this configuration will affect all jobs run in the entire CommCell.

It is also possible to override the default CommServe configuration for individual jobs by configuring retry settings when initiating the job. This configuration, however, will apply only to the specific job.

Note: Backup jobs for this Agent are resumed from the point-of-failure.

Configure Job Restartability at the CommServe Level

  1. From the CommCell Console ribbon, click the Home tab, and then click Control Panel.
  2. In the Data section, click Job Management.

    The Job Management dialog box is displayed.

  3. On the Job Restarts tab, select the job information.
    • In the Job Type box, select the job that you want set the restart capability for (for example, for DB2 backup jobs, select DB2 Data Protection).
    • Select the Restartable check box.
    • Optional: to change the maximum number of times that the Job Manager tries to restart a job, in the Max Restarts box, select the number of times.
    • Optional: to change the time interval between attempts for the Job Manager to restart the job, in the Restart Interval (Mins) box, select the number of minutes.
  4. Click OK.

Configure Job Restartability for an Individual Job

  1. From the Backup Options dialog box, click Advanced, then select the Job Retry tab and specify the following as desired:
    • Total Running Time - The maximum elapsed time, in hours and minutes, from the time that the job is created.
    • Number of Retries - The number of times that Job Manager will attempt to restart the job.
    • Kill Running Jobs When Total Running Time Expires - Option to kill the job when the specified Total Running Time has elapsed, even if its state is "Running".
  2. Click OK.

Controlling Jobs

The following controls are available for running jobs in the Job Controller window:

Suspend

Temporarily stops a job. A suspended job is not terminated; it can be restarted at a later time. Only preemptible jobs can be suspended.

Resume

Resumes a job and returns the status to Waiting, Pending, Queued, or Running depending on the availability of resources or the state of the operation windows and activity control settings.

Backup jobs for this Agent are resumed from the point-of-failure.

Kill

Terminates a job.

These controls can be applied to:

  • All jobs in the Job Controller.
  • All selected jobs in the Job Controller providing you have the correct security associations at the proper level for each job selected.
  • All data protection operations running for a particular client or client/agent.
  • All data protection operations running for a particular MediaAgent.

From the Job Controller of the CommCell Console, right-click the job and select Kill, Suspend, or Resume as desired.

  • When killing a job:

    Click Yes when the confirmation prompt appears if you are sure you want to kill the job. The job status may change to Kill Pending for a few moments while the operation completes. Once completed, the job status will change to Killed and it will be removed from the Job Controller window after five minutes.

  • When suspending a job:

    The job status may change to Suspend Pending for a few moments while the operation completes. The job status then changes to Suspended.

  • When resuming a job:

    As the Job Manager attempts to restart the job, the job status changes to Waiting, Pending, or Running.

Notes:

  • When backup jobs are suspended before the chunk size is committed, the job will restart the backup from the uncommitted chunk. Therefore, the backup size of a suspended job may be larger than the size of a normal backup job.
  • When a backup job is killed, you need to restart these Documentum jobs manually using Documentum software:
    • dm_DMClean
    • dm_DMFileScan

Additional Options

Several additional options are available to further refine your backup operations. The following table describes these options, as well as the steps for configuring them.

Be sure to read the overview material referenced for each feature prior to using them.

Option

Description

Startup Options

The start options are accessible from the Startup tab in the Advanced Backup Options window.

These options facilitate the startup behavior of the backup job.

The Default Priority option dictates the type of job that will be used in determining how the Job Manager will allocate resources for this job.

The Change Priority option specifies the priority for the job, between 0 (highest priority) and 999 (lowest priority). The Job Manager will use the priority setting when allocating the required resources. This is useful if you have jobs that are very important and must complete, or jobs that can be moved to a lower priority.

The Start up in Suspended State specifies that this job will start in the Job Controller in a suspended state and cannot run until the job is manually resumed using the Resume option. This option can be used to add a level of manual control when a job is started. For example, you could schedule jobs to start in the suspended state and then choose which scheduled jobs complete by resuming the operation started in the suspended state.

Refer to the Job Priority and Priority Precedence documentation for a comprehensive overview prior to using these options.

Job Retry Options

The job retry options are accessible from the Job Retry tab in the Advanced Backup Options window.

These options enable you to configure the retry behavior of the backup jobs.

  • Total Running Time - The maximum elapsed time, in hours and minutes, from the time that the job is created. When the specified maximum elapsed time is reached, as long as the job is in the "Running" state, it will continue; if the job is not in the "Running" state when the specified time is reached, Job Manager will kill the job.
  • Number of Retries - The number of times that Job Manager will attempt to restart the job. Once the maximum number of retry attempts has been reached, if the job has still not restarted successfully, Job Manager will kill the job. Note that this job-based setting will not be valid if restartability has been turned off in the Job Management Control Panel.
  • Kill Running Jobs When Total Running Time Expires - Option to kill the job when the specified Total Running Time has elapsed, even if its state is "Running". This option is available only if you have specified a Total Running Time.

Refer to the Job Management documentation for a comprehensive overview prior to using these features.

Start New Media

This option is accessible from the Media tab in the Advanced Backup Options window.

This option starts the backup/archive operation on a new media, which causes the following to occur:

  • If removable media is used, the current active media is marked as Appendable and a new media is used for the backup/archive.
  • If disk media is used, a new volume folder is created for the backup/archive.

If cleared, the operation automatically uses the current active media or volume.

This media management feature provides a degree of control over where the data physically resides, for example helping you to meet security or performance goals. This feature is useful in situations where you would like the data to reside on a new media, not appended to a media already containing backup/archive data.

For more information, see Creating an Exportable Media Set.

Another form of the Start New Media option is available from the library properties. Refer to the Start New Media section of the Library Properties documentation for a comprehensive overview.

Mark Media Full

This option is accessible from the Media tab in the Advanced Backup Options window.

This option marks media full, 2 minutes after the completion of the backup/archive operation. If any jobs are initiated within the 2 minutes, they are allowed to write to the media. If the job was associated with the prior media, new media (such as a new tape) will be used for subsequent jobs. (Applies to all backup/archive types.)

This media management feature provides a degree of control over where the data physically resides, for example helping you to meet security or performance goals. This  feature prevents any other data from being written to the media. If the job was associated with the prior media, new media (such as a new tape) will be used for subsequent jobs. (Applies to all backup/archive types.)

For more information, see Creating an Exportable Media Set.

Refer to the Export Media documentation for a comprehensive overview prior to using this feature.

Allow Other Schedules To Use Media Set

This option is accessible from the Media tab in the Advanced Backup Options window.

This option allows jobs that are part of a schedule policy or schedule and using a specific storage policy to start a new media and also prevent other jobs from writing to the set of media. It is available only when the Start New Media and Mark Media Full options are enabled, and can be used in the following situations:

  • When one scheduled job initiates several jobs and you only want to start new media on the first job.
  • When you want to target specific backups to a media, or a set of media if multiple streams are used.

For more information, see Creating an Exportable Media Set.

Data Path Options

The data path options can be accessed from the Data Path tab in the Advanced Backup Options window.

Data can be backed up to any compatible library and drive type in the CommCell. By default the system automatically identifies the data path for backup operations as configured in the CommCell. You can change this data path at the backup level by selecting a specific MediaAgent, library or drive pool in the Advanced Backup Options dialog. The data path change can be specified at the backup, schedule and the schedule policy levels. Note that the data path changes specified are only applicable only for the specific backup job/backup schedule.

Ensure that the specified MediaAgent and the library or drive pool are available and accessible, failing which the job will go into Pending state; failover protection does not apply here. Also, note that this setting overrides all other settings, including the data path settings at the Storage Policy, Subclient and the alternate data path (Gridstor).

The change data path options are useful in the following situations:

  • The preferred MediaAgent, library or drive pool is not available during the backup.
  • You wish to use a different MediaAgent, library or drive pool for load balancing purposes.
  • You wish to direct that backup to a particular location for compliance purposes.

Refer to Change Data Path documentation for a comprehensive overview prior to using this feature.

Vault Tracker

You can access the vault tracking options from the Vault Tracker tab in the Advanced Backup Options window.

Several Vault Tracking options are available to export and track media. These options are available only when a Vault Tracker license is available in the CommServe, and are only applicable for data protection operations using a storage policy associated with a library containing removable media (e.g., tape or stand-alone.)

Refer to Vault Tracker documentation for a comprehensive overview prior to using this feature.

Alerts

You can configure alerts for the specific backup job from the Alert tab in the Advanced Backup Options window.

Refer to Alerts documentation for comprehensive overview prior to using this feature.

Last modified: 3/1/2018 7:47:08 PM