Description

All about Backup Exec Catalogs

• What are Backup Exec catalogs?
• What is the difference between Inventory and Catalog?
• Where are the Backup Exec catalogs stored?
• What are the Contents of the Catalogs folder?
• Can Backup Exec Catalog Files be truncated automatically?
• What to do when a Backup Set is not available in the Restore Selection but Storage Media is
• What if the Tape is encrypted using a pass phrase and is being cataloged on a new Backup Exec Server?
• Backup Exec Catalog terminology.
• What is the correct way to run a Catalog Job?
• How to find the Family Sequence Number of a particular set
• Catalog Job Failures.
• Is backing up the Backup Exec Catalog Files helpful?  How is it done?
• Catalog - "Do Nots"
   

What are Backup Exec catalogs?

Backup Exec Catalogs are considered as the indexes that Backup Exec uses to ensure it can read the data that is stored in the Backup Media. The purpose of reading the data can be to restore data or duplicate a backup set to another secondary or tertiary storage.

When backup is taken on a media, the information about the set that is being written is also added to that storage media. This is referred to as Media Based Catalog which reside on the media. In order for Backup Exec to read the information about the set stored on the Backup Media, it needs to create Catalog files which it stores on Disk. These are called Disk Based Catalog (These are XML and FH files in catalogs folder of BE). If these files are lost they can be regenerated (Run a Catalog Job within BE) with the help of Media Based catalog as the media always contains the info about what sets are written on it until the media is overwritten.
 

What is the difference between Inventory and Catalog?

Inventory: This is the process of reading a media header and adding the information about that tape to the Backup Exec database.  "Media" is any medium where a backup is being written.  It can be BKF, OST, IMG and/or tape.

Catalog: This is the process of reading the metadata of the Backup Set from the storage media and then creating catalog files for Backup Exec (which BE would store in default Installation path\Catalogs folder).

The Backup Exec User Interface provides options to run Inventory and Catalog job as seperate Jobs. If both need to be triggered for a storage or a few media, then choose Inventory and Catalog now.

Where are the Backup Exec catalogs stored?

They are stored in BKF, IMG, OST and tape (media based catalog) and physically present at the Catalog location stored in the registry.  If the Catalog Files are not available in the Catalog location, BE cannot display the Backup Sets in the console.

The registry key is HKEY_LOCAL_MACHINE\SOFTWARE\Veritas\Backup Exec for Windows\Backup Exec\Engine\Misc\Catalog Path.

Note: The Catalog Path can be changed with BEutility.exe.  This utility is availablein the Backup Exec Installation Path. Select the All backup Exec Servers section from the utility, highlight and right click on the Backup Exec Server to launch the menu to run Set Job Log and Catalog Locations. Click "yes" on the prompt, if one comes up.  Select the checkbox to change the catalog location and paste or type the new catalog location path. Click "OK" and this should change the catalog location within BE and also move catalog files to the new location. To manually perform the same activity, refer to the Related Article linked below.

What are the Contents of the Catalogs folder?

Default Path - C:\Program Files\Veritas\Backup Exec\Catalogs
If Backup Exec is upgraded from Version before Backup Exec 16  - C:\Program Files\Veritas\Backup Exec\Catalogs

Subfolder with BE server name. There can be multiple folder with other BE server names if:

• This is a BE server with the Centralized Administration Server Feature Installed or it is a Managed Backup Exec Server.
• In an edge case there could be an old folder with a previous BE server name present. This can be the case if catalog movement from the old BE server name Folder to the new BE server name folder failed.  Contact Veritas Technical Support team to get these catalogs moved to the new BE server name Folder.  Backup Exec automatically adjusts the catalog entries in the Backup Exec database and also moves the catalog files to its new location (i.e. BE Install Path\Catalogs\<BackupNewHostnameFolder>).  If the automatic process which gets initiated when hostname of the server is changed is interrupted, the Catalogs\<OldBEServerNameFolder> may remain along with the new one.

Subfolder named Incompatible. Normally there are no catalog files here.

The important folder is the BE server name folder within Catalogs, as it contains all the files with Backup Set info:

GUID_n.XML
GUID_n.FH
GUID_n.Prev
GUID_n.SDR

The .XML files are the header files and contain the following Info:

1. Backup set date and time of backup,
2. GRT or non GRT Backup Set
3. Media cartridges used for the Backup, Backup media GUID etc. The XML file can be opened in a web browser or file viewer to check the backup set info.

The .FH files contain the info about what files are backed up to that Backup server.

The .Prev files are created when GRT enabled resources are selected in a Backup Job targetted to Disk and the Instant GRT section of the Job setting is set to use "Run a full catalog operation" as a separate job immediately after the Backup Job finishes.
These files are deleted after a successful run of delayed catalog.  If these files are present and no GRT Backup with the above option are running, then these files can be safely moved out of the Catalogs\servername folder.

The .SDR files are created by a successful SDR enabled backup job containing details needed for SDR recovery.

Can Backup Exec Catalog Files be truncated automatically?

The value of truncate catalog can be controlled from Backup Exec Icon (left top corner in UI) > Configuration and Settings > Backup Exec Settings > Catalog (Press F1 to open Help). It is recommended to keep the catalog till the backup media does not expire or is not needed. If this folder is lost, an Inventory and Catalog job will need to be initiated to read the storage media catalog content and recreate the catalog files from it. This can be a time consuming process, especially when a Backup Set is needed to be restored.  Hence, it is important to let Backup Exec manage these catalogs and keep the catalog files until the expiration of the Backup Set if possible.

Note: If the catalog files are set to truncate (by default, they are not) then only the FH files are deleted. The header files remain and a full restore can be performed.  For file folder level restore, a catalog job will need to be run. More information can be found in the Related Articles linked below.

The below figure shows how a backup set appears if the FH is truncated. Note that the icon next to the backup set from 03/11/2016 includes an exclamation point.  The entire set can be restored as the XML is present, but in order to restore individual items from that set, a catalog job will need to be run to regenerate the FH file.



What to do when a Backup Set is not available in the Restore Selection but Storage Media is

Run Inventory and Catalog on that Storage (or Storage Media) with Request all Media in the sequence for catalog operations and Use storage-based catalogs selected (These settings can be accessed by Clicking the BE Icon to the top left side of UI > Configuration and Settings > Backup Exec settings > Catalog)

Note: GRT, SDR, NDMP, True Image enabled Backup Sets should be cataloged with above settings. Use storage-based catalog should be selected.

What if the Tape is encrypted using a pass phrase and is being cataloged on a new Backup Exec Server?

This Alert can be seen when the pass phrase does not match what it was when the backup was taken:

The encryption key required by this set (key name: Key1) cannot be retrieved. To Catalog this set, click Tools, click Encryption Keys, and then create a key using the same pass phrase and encryption type as the original key.

You must close this message before you can go to the Tools menu. When running the Central Admin Server Option, keys must be created on the media server that is running the job. After you add the key, go to the Alerts pane, open this alert again, and then click Yes to retry operation for the set. Otherwise, you can click No to skip to the next set, or click Cancel to terminate the job.

It is important to note that the pass phrase is stored in the Backup Exec database in an encrypted format. It cannot be decrypted, even if requested due to security reasons. It is advised to provide the correct pass phrase and if that pass phrase is not available, the Backup Exec database of the original Backup Exec server where the backup was taken (which had the pass phrase configured) can be used to do the restore. The pass phrase should not be misplaced, or else the data cannot be decrypted and catalog or Restore of these Backup Sets will no longer be possible.

Note: Backup Exec configuration (Jobs, Catalog Info etc.) data is stored in two folders named Data and Catalogs Folder.  By default, these folders are located in the BE install path.  It is recommended to back up both these folders for ease of recovering Backup Exec configuration.

Backup Exec Catalog terminology

Media Family - When a Backup Job spans a media, these media will maintain the same Catalog GUID but the "{<GUID>}_n" substring will keep changing as more resources are added and appended to the tape media.  The Figure below shows an example of a backups spanning multiple tapes:

The below settings can be found at the Backup Exec button (in UI) by selecting Configuration and Settings, then clicking Backup Exec Settings > Catalog

Use storage-based catalogs: When cataloging a media, read the catalog information that is on the media. This allows the catalog operation to complete much faster, as it is not reading each file on the media to build a new catalog. This is how the Catalog job works if this option is selected. This can also be referred to as the FAST CATALOG METHOD.
If the Use storage media-based catalogs check box is not selected during a catalog operation, all files on the media must be read file-by-file to create a new catalog and this is a slower than if Use storage-based catalogs is selected. This can be referred to as the SLOW CATALOG METHOD.

Request all media in the sequence (MIS) for catalog operations:

• If selected along with Use media based catalog, the catalog process starts from the last media in media family and then works its way backwards till the first set of first media.
• If this option (MIS) is cleared, then use media based catalog remains selected (if it was selected earlier) and becomes greyed out.  In this case, if the Storage/Device is being cataloged, it will start from last media and work its way back.  If the admin wants to start the catalog from a particular media, then that particular media can be selected and the catalog will then start from this media. It will pick up the media you selected for the catalog job and then later pick up the last media in the media family for catalog operation and work its way backwards (last media used by backup to first media used by backup).
• If request media in sequence is cleared along with use storage based catalog, then the catalog job starts cataloging from the first media or the media which was selected to start the catalog operation, then moves from lowest used media to the last media in media family.

Note: In all of these cases, ensure to Inventory all the media which are being cataloged.

Family ID: Backup media of individual resources that span media have the same Family ID. This helps to know how many media a particular resource backup has spanned.  Inventory Job History can be checked by Family ID to group all media of a particular set. This is the case with Disk Storage.

On Tape, Family ID is maintained and all resources backed up to a particular tape will maintain the same family ID. The last set which is written to the tape, if it spans to the next tape, the next tape media will also maintain the same family ID. On tape, the resource backups are appended (if not specified to overwrite i.e. start from beginning of the tape) which is different from how the backups are done on disk. On disk, media spans for a backup of particular resource (if backup is unable to fit on one media) but as soon as the backup of this particular resource completes then a new fresh disk BKF media is created to start the backup of the next resource. So, on disk only, one set may be affected if one of the backup media is missing, but on Tape, one or many backups can be written to that tape (until it is full and if the backup spans tapes, they will carry the same family ID) and the catalog job will report an error that the media sequence is broken.  For this reason, it is important to ensure all tapes of the backup sets are available; otherwise, the catalog job may report an error.  More importantly, data will only be restored from the available tapes. Utmost care should be taken with resources which are just one large file and may span to another media (for example: a database backup).

What is the correct way to run a Catalog Job?

An Inventory of all the media which constitutes the Backup set is important to correctly catalog.  Inventory jobs read and record the info about the media (media header, Media Label etc.) to the Backup Exec database. Some important info that Backup Exec gets from Inventory Job:

It helps to understand the family sequence number and which Medias are part of same media family etc. This info is very helpful in troubleshooting catalog related Issues.

How to find the Family Sequence Number of a particular set

The Family Sequence Number is the order in which the media is picked and used by Backup Exec during a backup operation.

For Backup Tape Media, there can be multiple sets on one tape media and sets can span to other tapes, in which case they will carry the same Media Family ID.

For Backup Disk Media, each backup set will have a unique Family ID.  If more than one media gets used to back up a resource, then all these media will share the same Family ID.  The Job History of an Inventory Job will show the details about media. Media can be filtered with the Family ID and the sequence in which they were picked can be extracted from this detail.  It is beneficial to know this in case a catalog job fails due to a missing media.

Catalog Job Creation Error

Unable to create the Catalog Items. Unknown reason: 00000057
Refer to the Related Article linked below.

Catalog Job Failures

1. Final error: 0xe00084ca - The data being read from the media is inconsistent
Final error category: Backup Media Errors
Catalog
An inconsistency was encountered on the storage media in Disk storage 0001.
V-79-57344-33994 - The data being read from the media is inconsistent.

Note the following details in the Job Log:

• Did the catalog job fail on several media or only on one specific media which has some inconsistency? This can be easily found by reviewing the failed catalog job. Check the places in the job log where the error is seen.
  Just before this error in the Job Log, check the media which was mounted. Normally, this problem is seen when a media contains an abruptly ended backup set.  It can happen that a backup job was aborted midway by ending the bengine.exe process, resulting in the end of the set (end marker) info not getting written. These media would normally show as Not Appendable (End Marker Unreadable) in the appendable until column of Storage >Media tab (ShowHiddenMedia needs to be enabled to review this - please refer to the Related Article linked below). The errors will be received as the issue occurred during backup time, hence the error can be regarded as a message to inform the admin of an inconsistent set of Media. It may also happen that there is a problem with the media which can be determined by checking if there are any hard read/write errors on Media or on the Drive where storage is configured. For Disk, always check the system and application event logs for any NTFS errors.
• Take note of the device name where the catalog was run (Eg: Disk Storage 0001 in the case seen above).  A chkdsk can be run to ensure the drive file system is fine, just in case any NTFS errors are seen on the disk holding the disk storage.

Job Log excerpt:

Drive and media mount requested: 9/24/2016 12:38:39 AM \* Next media requested *\
-Device and Media Information
Drive and media information from media mount: 9/24/2016 12:39:41 AM \* Media is now mounted *\
Drive Name: Disk storage 0001 \* Device Name *\
Media Label: B2D000128 \* Media Labe which was mounted. In this case this is the one which is cataloged and throws the error *\
Media GUID: {eefc9bab-90aa-4a88-91a9-c60d5e5caec6} \*This is the media GUID for above B2D000128 Media. Mark that you may have several device and same media label but BE determines the correct media by looking at this media GUID *\
-Job Operation - Catalog
Loading media-based catalog set list from the storage media. \* When you read Loading Media-based catalog list from the storage media means we are reading catalog section from the media i.e. use storage media based is selected and it is called fast catalog *\
An error was encountered while attempting to read the Media Based Catalog's Set Map data. \* fast catalog had an error *\
Attempting to catalog the media by reading the data area of each backup set. \*This is when we switch to reading that media using slow catalog method. We read the data area and then create the catalog files for the set we can find *\
-Set Information - \\WIN-AD.BACKUPEXEC.LOCAL\System?State \\WIN-AD.BACKUPEXEC.LOCAL\System?State
Backup of \\WIN-AD.BACKUPEXEC.LOCAL\System?State \\WIN-AD.BACKUPEXEC.LOCAL\System?State \* resource being cataloged *\
Backup set #1 on storage media #1
Backup set description: ""
This backup set has already been cataloged.
Processed 2 bytes in 1 second.
An inconsistency was encountered on the storage media in Disk storage 0001.
V-79-57344-33994 - The data being read from the media is inconsistent. \* error *\

2. Error: 0xe0000900 - The requested media is not listed in the media index and could not be mounted. To add the media's catalog information to the disk-based catalogs, run an inventory operation on the media and resubmit the Catalog operation.

The requested media (sequence number 3 in the media family being cataloged) was not listed in the catalog's media index, and therefore it could not be mounted.  No data from this media will be included in the disk based catalogs. Performing an Inventory operation on this media and rerunning the Catalog operation should resolve the problem

As the error suggests, the media was not found during the catalog operation. It is possible that the media info does not exist in the database. The reason why the info about a particular media may not exist in the BE Database could be because the media does not exist anymore (was overwritten) or was not put in the storage.  It is then important to review the Job log and check the place where the error is seen.  Then, review the Inventory Job History Log to identify which media is missing. If the media is found, then put the media back in storage, start Inventory and Catalog.

Note 1: The Catalog Job Log does provide a hint about which media is missing. For example in the above case, it says "The requested Media and then it mentions sequence number 3 in media family". One can easily find out the backup set that is missing media. If the media is not found then only data in the available media can be restored. It is always recommended to perform the backup with use media based and use media in sequence checkbox selected from Global Backup Exec Settings > Catalog.

Note 2: The same error is reported if the Catalog of Media where the backup set spanned, but next media was not provided and the backup job was cancelled for lack of tapes.  The error is reported on the last set it is cataloging, as Backup Exec thinks it spanned but cannot find the next tape media as it was never inserted.  This issue is more likely to be found with Tape media. On Disk, this issue can occur as well, but only when the next media to span the set does not get created due to space filling up on the disk storage and job is cancelled.

Is backing up the Backup Exec Catalog Files helpful?  How is it done?

Yes, from a Disaster Recovery perspective, the Data and Catalog folder present in the BE Install Path (or where they exist in case these are kept at different locations) of the BE server should be backed up.

It is most helpful when a Disaster Recovery situation arises in the data center and all servers need to be restored.  If a Backup Exec server is restored to its pre-disaster state, then the time to catalog the media is saved which helps allow for a faster recovery of other servers in the environment.  For this reason, a Full backup of the Data and Catalog folder is recommended every day. This is the bare minimum required to restore Backup Exec Jobs, catalog Info, settings etc.

Note: This is the default path where Data and Catalog exist.

Catalog "Do Nots"

• Do not rename or delete the Backup Exec Catalog Folder if the Catalog Job fails. The information about earlier sets will be lost. If Catalog files are deleted, earlier backup media will need to be recataloged to regenerate the catalog files.
• Backup Exec is now integrated with Veritas Information Map. Metadata from the catalogs are sent to InfoMap for IM to be able to understand the type, age etc. of the Data that exists from various Data sources (Backup Exec being one of them). This is one of the secondary reasons to not delete or lose the catalog folder.
• The media (especially disk media) should not be deleted from Disk physically - for example, if the Disk storage is configured on the F:\ drive, then the BKFs should not be manually deleted from F:\. Backup Exec incorporates a process called Data Lifecycle Management (see Related Article) which clears the Disk Media and catalog when the set expires. If automatic deletion of backup sets are not working, please contact Veritas Technical Support.