MEDIA includes many features designed to simplify and automate all aspects of media management. Some of the most important highlights are listed below:

A comprehensive, on-line media database...

• System maintained on-line database: describing the off-line media and their contents. This allows the user to search for specific files without the need to mount the media. Simple commands allow the user to search directories and selectively issue commands to read and write files to or from the selected medium.
• Directory listings of data in database: Using the MCL interface both media contents and media database information can be examined. For DCL users, the DIRECTORY command can be used to review all data in the database. Using wildcards, it is possible to search the media directory to locate a specific file and medium. The format of the DIRECTORY command is similar to the VMS DIRECTORY command, including such qualifiers as /EXCLUDE, /SINCE and /BEFORE.
• Directory update procedures: The procedure to extract the directory information from the off-line media and add it into the on-line library database is referred to in this manual as UPDATE (or SUBMIT UPDATE for MCL users). This procedure is performed at the end of any media job. Several update procedures are available in the standard product distribution. These include VAX TAPES, IBM TAPES, DOS-11 TAPES, UNLABELED TAPES and VAX DISKS. For media recorded in other formats, the necessary interface exists to allow a local program or command procedure to be supplied.
• Automated backups: The BCKMGR and MEDIA software is designed to allow for creation of backup jobs using a simple command file. When executed, such a file would select required media, exercise the DEC backup utility and then properly update the Media database to reflect the results of the backup process. ¹
• Automatically created batch jobs: Batch jobs are automatically created for any command which requires the medium to actually be mounted on the device. This streamlines accessing of the drive and enables an orderly sequence of mount requests to the OPERATOR.
• Record definitions: All the record formats used in the library database are supplied in the documentation. This readily allows such programs as Datatrieve to extract and create custom reports based on data in the Media Library.
• Database Information: The database contains such information as:
  
  • The physical attributes of the off-line media, such as size, density, reel length, media type (TAPE, RP06 ...), error count, access count, etc.
  • Several dates, such as added on, first used, last used, etc.
  • Storage capacity of the media including amount used, in feet for tapes and blocks for disks.
  • Conditions to use during mounting, updating and validating the media.
  • All available RMS attributes of files contained on the media.
  • Many supplemental locations such as storage location, comment field, protection setting, associated volumes, etc.

• Cataloged and INCORE media: There are two distinct types of media that are recognized by the Librarian. The cataloged variety are those that are permanently contained in the library database. An INCORE medium is one that has been temporarily cataloged by a user. Any INCORE media will be removed whenever the user logs outs. The INCORE type is very useful in the case where a user brings in a tape that he wants to process once and then takes it with him.
• Label validation: Several different types of labeling schemes are recognized by the Librarian. They are IBM standard labels (tapes), ANSI standard labels (tapes), FILES-11 labels (tapes/disks) and DOS-11 labels (tapes).
• Nonstandard tapes: Several facilities and commands are supplied to allow nonstandard formated tapes to be incorporated and processed by the Media System.
• Volume sets: Media can be bound together as a volume set. Several commands exist to attach or detach a medium to or from a volume set. The OPERATOR is allowed to extend a volume as needed during a job execution.
• On-line contents file: The information about the contents of an off-line medium is kept in a contents file. These contents files can be in any one of three formats. The native format, called SDR, is generated by the included update procedures. The other two formats can be added into the library database by the user. When using the VMS BACKUP command, the journal file can be used as the contents file. These are referred to as BJL files. The third variety is a sequential text file format. It is called TDR.
• Shadow media: In many cases, it is important to keep two copies of the same information. A shadow medium is the second media. Once a shadow medium has been assigned, any command that would write to the master volume will produce a similar operation for the shadow volume. ²
  

A detached server process...

• Device control: The Librarian uses a detached server process to monitor and control the allocation of removable tape and disk drives. Only those drives listed in the controlling parameter file will be affected. Drives are allocated on a least recently used basis. Independent jobs using the same medium will automatically execute without any OPERATOR assistance to remount the medium.
• Error logging: The server process monitors the error activity on the system. Any errors that are associated with a known medium will be added into the error counter kept for that medium. This error count can be very useful in determining when a particular medium should be replaced or re-certified.
• Device allocation control: In many cases, it is desirable to force all removable drive allocations to be processed by the librarian system. This can be done by requesting the server process to allocate all the defined drives and only turning them over to proper and validated requests via the librarian. This will disable the VMS ALLOCATE command and only allow the MEDIA ALLOCATE command.

Easy to manage media pools...

• Pool management: All media are classified into one of three pools: AVAILABLE - those that can be allocated by a user on request; RELEASED - those that have been returned to the system but not yet made available by the Media Librarian; ALLOCATED - those that are currently in use and assigned to a specific user.
• User allocations: Media can be easily allocated from the available pool, one at a time or in groups, via a simple command.
• Automatic migration: Support software is provided for automatic recognition of expiration dates. Media can be automatically moved from the allocated pool to the released pool after expiration. After a holding period, they can be automatically moved into the available pool.

Cluster and Network support...

• Clusters: Full support is provided for any combination of VAX CPU's in a cluster. Jobs will automatically be placed in the right queue for the node that has the correct type of tape/disk drive.
• DECnet: Full support is provided for any degree of sharing across a network. Geographically distributed media libraries are supported while maintaining a central database. Jobs will not be created at one location that entail reading a tape stored at another distant location. All users of the Librarian have full access to the entire collection.

User dependent media access and security...

• User versus operator added media: Whenever a medium is cataloged into the library database, it is noted what type of user added the media. Three classes of user are recognized: OPERATORS, MANAGERS and general users. A general user can only delete media that he has added to the database. An OPERATOR can delete any type of media.
• Access controls: With each medium cataloged into the library database, a UIC-based, access control string is kept. This access list is grouped in SYSTEM, OWNER, GROUP and WORLD users. Each group has an associated access specification.
• Volume security: All volumes initialized by the librarian are set to be owned by the UIC and protected as indicated in the on-line database. This will effectively (except for the owner of the reel) disable the VMS MOUNT command for a user. The MEDIA MOUNT command will mount such media and set the parameters once mounted to match that defined in the library database. The Librarian goes to great lengths to prevent accidental destruction of data on media. Before any operation is allowed on a media, the label is checked to verify that it is indeed the correct media.

1.2 Restore system

VAULT, Restore via search and Restore entire disk makes the restoring of lost or deleted files easier by allowing restoration from remote nodes and even remote PC's. The process is further simplified by the ease of use of the MOTIF Work Center .

1. VAULT Saveset Manager allows users to easily identify and restore any files that have been copied onto off-line storage during normal archive, incremental and full backups.
   VAULT maintains a directory structured database that lists information about files that have been backed up. The main objective is to facilitate identifying what has been backed up and simplify the selective file restore process.
   The VAULT database contains an entry for every file. This entry gives the file storage location, when backed up and other attributes. Each backup can be individually set to add or not to add information to the database. The major features of VAULT are:
   An on-line directory database...
   
   • On-line directory: A directory structure which parallels the on-line directory is maintained for all files that are in the backup system. Just as the VMS DIRECTORY command gives a listing of the on-line files, the VAULT command gives a listing of all the off-line files which have been backed up.
   • File protection: A user can only view or request a restoration of files that he has access to. The UIC based protection attributes are maintained in the VAULT database.
   • Unloading of unused files: Permanent file unloading can easily be achieved by using BCKMGR to create a regular job that copies old files to tape. This is best done at least twice. Once both tape jobs have executed, a third job (again automatically done by BCKMGR) updates the VAULT database and deletes the files. Simple backup and restore...
     • Backups: Regular backups, either incremental or full, are easily executed using the BCKMGR system. After each backup is executed, the VAULT database is updated. From that moment on, all users can easily identify and selectively restore any file that exists on a backup tape. All backup tapes and files are kept using VMS BACKUP thus allowing fully portable tape formats and compatability with all other VAX sites.
     • Simple restore: Any file found in the VAULT database will be restored. A RESTORE request will automatically create a MEDIA job. The command sequence of the batch job will request the OPERATOR to mount the proper tape or disk then restore the specified files.
2. Restore entire disk: The Restore entire disk program can be used to restore a disk because of errors, disk crash, relocation or any other reason. The latest full disk backup and latest incremental backups are combined to restore the disk to it's most recent state.
   Restore entire disk determines the volumes required for the recovery and restores up to the last backup that occured. The user may designate the target drive in case the recovery will apply to a different drive.
3. Restore via search: The Restore via search facility allows the OPERATOR to create a detailed search criteria to locate needed files. By selecting such specific criteria as files names, saveset names and backup job, it is possible to locate files on any volume.

BCKMGR allows users to create and submit backup jobs at regular intervals while designating the source disk and offline media for the job. Virtually any non backup job can be submitted as well using a list of DCL commands. The major features of BCKMGR are:

Complete control of job characteristics...

• Job attributes: The characteristics of a job can be designated in detail, including the number of media required, submission options, history options and access controls. In addition, the priority and status of the job may be monitored after it is submitted.
• Source disk assignment: The disks to be backed up can easily be assigned using a simple list of devices in MCL or DCL. These devices will be backed up when the job is executed.
• Target media assignment: The media to be used for a backup job may be explicitly assigned or deassigned and from that point on dedicated exclusively for that purpose. The media then belongs to the BCKMGR database. The quantity needed for the job is based on the type of media allocation and deallocation used. The three choices are CYCLE, GENERATIONS or ONCE.
• DCL commands and parameters for non-backup jobs: If the job to be created is not a backup job, a list of DCL commands entered into an editor will be performed upon execution of the job. Parameters can also be designated in place of the device designation that takes place for a backup job.

Auto submit cycles and full disk recovery...

• Auto submit cycles: Whether or not SCHEDULE is installed on the system, the job you create can be submitted automatically using predetermined intervals. By simply giving the job name, start time and interval, the job will run according to that interval. In addition, if SCHEDULE is installed on the system, the job may be linked to a SCHEDULE job and run that job using the auto submit cycle.

1.4 Backup Modes

• allow for recovery from a disaster
• provide for audits and legal requirements
• free up space on the on-line disk

The type of backup job is specified by the MANAGER when he creates the backup job. At most sites, the MANAGER will set up daily incremental backups and weekly and/or monthly full backups. Unloading backups might be scheduled daily, weekly, monthly or yearly depending on the site. Check with your MANAGER to learn the schedule of backups at your site.

In addition to standard Full and Incremental backups of on-line media, backups may be performed on remote nodes incorporating VMS clusters, PC networks and a variety of servers all to a central VMS device. Agent NT and Agent UNIX backup modes enable backups of remote nodes to remote backup devices.

Each BACKUP job in the BCKMGR system is defined by a set of attributes that describes what is to be backed up and how to do the back up. These attributes can be grouped as follows:

• What disk or files to back up. This is the parameter list. These disks or files can be on the local system or on a remote system.
• How to do the back up and conditions to examine. This is the backup mode and optional job settings.
• How many and what type of removable media to use. This is the media type, counts and flags.
• Other control attributes. This is the job log setting, protection codes and user name.

One of the key attributes is the backup mode number. The backup mode controls how the backup operation is performed. Several predefined modes are provided with the system. Others can easily be added by the user as needed.

Table 1-1 below lists the various predefined mode numbers.

Table 1-1 Backup modes

	Standard		Scratch
Backup Type	Incremental	Full	Incremental	Full
On-line	2	3	4	5
Appending	13	14	15	16
Remote VMS	9	10	11	12
Remote PC	17	18	19	20
Remote NFS	21	22	23	24
Agent NT	25	25	25	25
Agent UNIX	26	26	26	26
Backup Type	Incremental	Full	After Image Journal
RDB	6	7	8