Previous | Contents | Index |
A typical OPCOM mount request for a second volume where the second volume is known to the system appears as:
%%%%%%%%%%%% OPCOM 1-APR-1996 10:00:03.33 %%%%%%%%%%%% Request 34, from user SMITH MOUNT relative volume 2 (1202) on MTA1: |
In this case you just mount the requested reel and the system will proceed.
A typical OPCOM mount request for a second volume where the second volume is UNKNOWN to the system appears as:
%%%%%%%%%%%% OPCOM 1-APR-1996 10:00:03.33 %%%%%%%%%%%% Request 34, from user SMITH MOUNT new relative volume 2 (120002) on MTA1: |
In the above sample the first volume was 1200. VMS modified the name by appending the digits 02 to the 1200. 19 Here you must attach a new volume and reply correctly to the request. The correct response to this request is:
$ MEDIA UNKNOWN ATTACH_TO/LOG 1200;0 1210 allocated 1210 attached to 1200;0 |
$ REPLY/INITIALIZE_TAPE=34 "''MEDIA_INT'" or $ REPLY/BLANK_TAPE=34 "''MEDIA_INT'" |
If the requested volume has been identified in a previous OPERATOR request, just perform step two above. The DCL SYMBOL MEDIA_INT, after the execution of the UNKNOWN medium allocation, will contain the known internally recorded label of the tape. The REPLY/INITIALIZE will check all the access rights before initializing the tape. The REPLY/BLANK will assume that the tape is blank and just write a label. 20
2.5.2 Extending a BACKUP tape
The OPERATOR requests covered in this section are those that are
generated by media jobs not created via the SCHEDULE system as
described in the Manager's Guide Section. 21
The VMS BACKUP program does all volume switching and labeling internally. The labels of all volumes that need to be used are passed to it via the /LABEL qualifier. When a backup overflows onto the second and later media the behavior is similar to the previous case of ANSI volumes.
A typical OPCOM 22 mount request for a second volume where the second volume is known to the system appears as:
%%%%%%%%%%% OPCOM 15-OCT-1996 14:49:59.49 %%%%%%%%%%% Request 4, from user SYSTEM on ISE Please mount volume 750 in device _ISE$MUA0: BACKUP requests: Saveset ISE1.BCK, Volume number 02, write ENABLED |
In this case you just mount the request reel and the system will proceed.
A typical OPCOM mount request for a second volume where the second volume is UNKNOWN to the system appears as:
%%%%%%%%%%% OPCOM 15-OCT-1996 14:49:59.49 %%%%%%%%%%% Request 4, from user SYSTEM on ISE Please mount volume 750_02 in device _ISE$MUA0: BACKUP requests: Saveset ISE1.BCK, Volume number 02, write ENABLED |
In the above sample the first volume was 750. VMS modified the name by appending _02 to the 750. Here you must attach a new volume and reply correctly to the request. The correct response to this request is:
$ MEDIA UNKNOWN SET/INTERNAL=750_02/STATUS=MOD_LABEL 756 allocated |
$ MEDIA 'MEDIA_EXT' ATTACH_TO 750;0 756;0 attached to 750;0 |
19 VMS always modifies positions 5 and 6 of the 6 character volume label that is placed on a magnetic tape. The last two characters are set equal to the volume number in the volume set.20 The VOLPRO privilege is needed to use the /BLANK qualifier.21 Standard backups created via the BCKMGR_MAINT.COM procedure was discussed on Section 2.4.1.22 These messages were generated on a VMS V5.2 system. |
This chapter is designed for those people who are assigned the responsibility of setting up and overall management of the MEDIA Librarian. Most questions come from the management of off-line tapes and disks. Therefore, this section concentrates on problems dealing with the MEDIA Librarian.
This chapter also features a description of how to set up MEDIA including the fundamentals of configuration and a description of DCL configuration.
To be a Media Manager you need the MANAGER privilege.1
1 The BYPASS privilege or a system level UIC will similarly grant you system manager access. |
The concepts mentioned in Section 2.1 are applicable to MANAGERS as
well as operators. By reading that section first, the following
concepts will be easier to understand.
3.1.1 TYPE category
Pools can be further subdivided into subpools by defining TYPE categories. The TYPE is a ten-character field in the MEDIA DATABASE that can be any alphanumeric word. This allows for
You can subdivide your media library by physical characteristics (i.e., nine-track tapes, TK50 and TU58 tape cartridges and disks). You can also subdivide the media based on function (i.e. USERTAPE and SYSTAPE). 2 DEVICE parameter
You can use the TYPE category to associate a type with a device, so that whenever a user requests a device allocation for a medium, the correct type of drive is always allocated (i.e. a USERTAPE can be used on MTA0:, MUA0: or MFA0:).
Again, this subdivision is invisible to users and operators.
MEDIA tracks the type categories that you define and makes
sure that the appropriate medium is mounted on an appropriate drive.
Jobs are submitted to an appropriate queue.
3.1.2 Volume sets
MEDIA can be attached together, so that the system considers them to be a single medium with several volumes. This can be very useful for organizational purposes.
While attached, the entire volume set can be addressed by the name of
the first volume in the set, although the individual media do retain
their separate names. A medium must be detached from the set before it
can be released or independently overwritten. There are several
commands to attach or detach a medium to or from a volume set. It is
also possible for the operator to extend a volume as needed during a
job execution.
3.1.3 The Media server
The server process is a detached process that is continuously running on the system. The behavior of the MEDIA Server is controlled by a set of parameters. These parameters define the devices that MEDIA manages, the location of data files, the batch queues used for MEDIA jobs and default settings for a number of activities. The server uses the following files:
File name | Contents |
---|---|
MEDIA_LIBRARY:MEDIAS.EXE | Server executable image |
MEDIA_LIBRARY:MEDIAS.LOG | Log file of all server events |
SYS$MANAGER:MEDIA.VAR | Parameter file |
SYS$COMMON:[SYSMGR]MEDIA2.VAR | Optional cluster common parameter file |
MEDIA_LIBRARY:MEDIAS.DAT | Across cluster device coordination data |
It monitors drive activities, creates batch jobs as scheduled, coordinates all medium access, and performs various related control functions. For example, it controls the allocations of drives on a least recently used basis. Only those drives listed in the controlling parameter file are affected.
The server maintains a log file that contains a list of all server events: allocations, mounts, dismounts, deallocations, job initiations and job completions. It contains a listing of any error messages generated while performing these functions.
If you experience odd or erratic behavior by server, look into this log
file for any error messages.
3.1.4 System configurations
Your system may be composed of:
MEDIA
supports any of these configurations.
3.1.4.1 Networked systems
Full support is provided for any degree of sharing across a network of multiple CPUs, each with its own associated drives. MEDIA maintains a central DATABASE. MEDIA ensures that jobs are not created at one site that requires reading a medium stored at another location. 3
3.1.4.2 OpenVMS clusters
There are three kinds of OpenVMS cluster systems, based on how the
resources are shared:
Full support is provided for any combination of VAX CPUs in a cluster. Any degree of sharing or separation can be maintained. Jobs are automatically placed in the right queue for the node with the correct type of tape or disk drive.4
3.1.5 The data files
Each of the three main application programs in MEDIA keeps
data files that make up the database for that application. These files
are independent of each other and can be repositioned separately. To
accommodate various hardware configurations, the system has a variety
of controls to vary the location of the database files. These are
discussed in Section 3.3.2.1.
3.1.5.1 MEDIA data files
The MEDIA DATABASE consists of the following files:
File name | Contents |
---|---|
MEDIA_LIBRARY:MEDIA.DAT | Main control file, contains header information. |
MEDIA_LIBRARY:*.BJL | Contents information file, in BACKUP journal format. |
MEDIA_LIBRARY:*.SDR | Contents information file, in expanded sequential format (MEDIA's internal format). |
MEDIA_LIBRARY:*.TDR | Contents information file, in a sequential text file format. |
There is usually only one MEDIA library per data center and it
is shared across the cluster or network. It is an indexed file keyed by
the medium external id. All of the data files for the MEDIA
DATABASE are located using the logical name MEDIA_LIBRARY. In a
networked environment, you can reposition the relevant data files using
the parameter CENTRAL 5
3.1.5.2 BCKMGR data files
The BCKMGR DATABASE consists of the following files:
File name | Contents |
---|---|
BCKMGR.DAT | Main backup job data file |
BCK_*.DAT | DCL command files for jobs containing more then 5 commands. |
There is usually one BCKMGR DATABASE per batch queue structure.
There is only one of these files in a cluster and thus only one
BCKMGR database should exist. A networked node would have its
own queue file and thus it should have its own BCKMGR
database. The BCKMGR.DAT data file is an indexed file keyed by the job
name. The parameter BGR_CENTRAL can be used to redirect the location of
these files to any other directory other than
MEDIA$LIBRARY:.
3.1.5.3 VAULT data files
The VAULT database consists of a group of ADR files.
File name | Contents |
---|---|
MEDIA_LIBRARY:000000000.ADR | Count file. Contains the number of the last ADR created. |
MEDIA_LIBRARY:000000001.ADR | Device level dir ectory file. |
MEDIA_LIBRARY:NNNNNNNNN.ADR | Directory file. Each one contains the file information from one DIR file. |
These files are interconnected in a tree like fashion. This is identical to the way on-line directories and sub-directories are connected.
The placement of these files can be controlled by the use of the VDISK
parameter. 6
3.1.6 The executable files
The executable and related files used by the system can be placed
anywhere. The locations shown below are those set by the installation
procedure.
3.1.6.1 MEDIA executable files
The files for MEDIA are:
File name | Contents |
---|---|
MEDIA_LIBRARY:MEDIA.EXE | Main command executable |
MEDIA$LIBRARY:MEDIAS.EXE | Server executable |
MEDIA_LIBRARY:FORMAT.EXE | FORMAT utility executable |
MEDIA_LIBRARY:MEDIASHR.EXE | Call interface sharable image |
The files for BCKMGR are:
File name | Contents |
---|---|
MEDIA_LIBRARY:BCKMGR.EXE | Main command executable |
MEDIA_LIBRARY:BCKMGR.HLB | Help library file |
The file for VAULT is:
File name | Contents |
---|---|
MEDIA$LIBRARY:VAULT.EXE | Main command executab le. |
2 Any new TYPE category must be defined to the system by the QUEUE and DEVICE parameters before it can be used.3 This is done by using different TYPE words at each location.4 All nodes in a cluster must share MEDIAS.DAT even if the cluster is heterogeneous5 MEDIA replaces the logical name MEDIA_LIBRARY: with the CENTRAL parameter value before opening the file. On a remote node it is important to have this parameter contain both a node name and an access string i.e. node``user password''::disk:[directory]. so that they are shared across the whole network.6 The VDISK parameter allows all the ADR files related to one on-line disk to be placed in a selected directory on any on-line disk. The BCKMGR_MAINT.COM procedure automatically sets up these parameters to point to subdirectories of the MEDIA_LIBRARY location. |
Previous | Next | Contents | Index |