BACKUP-SYSTEM

The BACKUP SYSTEM is a subsystem used for copying files from one storage medium to another. BACKUP means making an extra copy of a disk pack or individual files which can be kept in a safe place. If the need arises, this copy can be used later for recovery of deleted files or corrupted data. This manual describes: BACKUP-SYSTEM ND 210337H, the stand-alone programs for taking backup, and the SINTRAN III commands for taking backup.

The BACKUP-SYSTEM offers a variety of facilities for copying files to and from disks, floppy disks, and magnetic tapes. Files stored on remote computer systems may also be copied.

The files may be copied for archiving, backup, or other purposes. To enable communication with other computer installations, the ANSI standard label format is available for magnetic tapes.

Entering commands to the BACKUP-SYSTEM is easy, but slightly different from SINTRAN III. Some commands have subcommands, i.e., the parameter sequence is not solely determined by the first command entered. Online help information is available for every prompted command, subcommand or parameter at all levels of communication.

The old SINTRAN III commands @COPY-USERS-FILES, @CREATE-VOLUME and @LIST-VOLUME are now available under the BACKUP-SYSTEM. The commands have some extended and altered facilities.

The BACKUP-SYSTEM can handle files produced under older versions of SINTRAN III.

This article is a stub. You can improve NDWiki by expanding it.

Storage media for backup

The storage media for backup copies of files are hard disks, floppy disks, magnetic tapes or streamer tapes.

Organisation of files on storage media

Depending on how files are organised on storage media they are said to reside on a DIRECTORY or on a VOLUME. Files stored on hard disks must be stored on a DIRECTORY, files on floppy disks can either be stored on a DIRECTORY or on a VOLUME whereas files stored on magnetic tape will only be stored on a VOLUME.

Label format on magnetic tape volumes

Implementation of magnetic tape volumes is based upon the American National Standard Magnetic Tape Labels for Information Interchange X3.27-1969.

However, some deviations from the standard have been made. Deviations are marked by a dollar sign ($) in the explanation.

General rules

- The general tape layout is as follows:

                VOL1 HDR1 HDR2 UHL1*-file-*EOF1
                                *HDR1 HDR2 UHL1*-file2- * ( EOF1 OR EOV1 ) **

where VOL1,HDR1,HDR2,UHL1, EOF1, and EOV1 are tape labels, and asterisks are tape marks.

- All labels are 80 character blocks.

- All information in the labels is recorded as ASCII characters with the parity bit cleared. All unused character positions will contain spaces. $ The user option field (3) in the label UHL1 contains binary information.

- File data is recorded as I * 2048 character, blocks where I is an integer from 1 to 8. These blocks may contain any character (0-255).

$$$$ Deviation From Standard

- Only the first file on a volume may be extended to other volumes.

A nonstandard label, HOLE, has been introduced. This label can be inserted between the file data blocks. The important information in this label is a 32-bit binary number contained in characters 77-80 of the label. The BACKUP-SYSTEM uses this number in the following way:

Each 2048 character block on the tape corresponds to a 1024 16-bit word block on the disk, referred to as a page. The pages are numbered O, 1, 2, 3, etc., to establish a logical sequence of pages. If the logical sequences are not continuous, then a HOLE label defines where the next block on the magnetic tape logically belongs in the disk file. In order to represent a logical HOLE on the magnetic tape, the HOLE label will be inserted in front of the next block, stating this block's logical number. Blocks of 2048 characters without a HOLE label are expected to belong to a continuous logical area, and will cause the logical block number to be incremented by one.

Example:

log. block no: O          5    6    7        100  101       120
              data HOLE data data data HOLE data data HOLE data
                   (5)                 (100)          (120)

where data represents file data blocks of 204 characters, and HOLEs are labels. The contents of the HOLE labels are shown in parentheses.

Labels

Volume header label

                    VOLUME HEADER LABEL
POSITION FIELD NAME                LENGTH CONTENTS
1-3      1     label identifier      3    VOL
4        2     label number          1    1
5-10     3     volume serial number  6    (volume name) $
11       4     accessibility         1    (space)
12-31    5     (not used)           20    (spaces)
32-37    6     (not used)            6    (spaces)
38-51    7     owner identification 14    (name of owner) $
52-79    8     (not used)           28    (spaces)
80       9     label standard level  1    (spaces)

$ field 3 and 7

- These fields may contain any alphanumeric characters. If the field is not fully filled with characters, the last character in the string is an apostrophe. This character is used to mark the end of the string and is not part of the name. The unused part of such a field is filled with spaces.

First file header label

               FIRST FILE HEADER LABEL
POSITION FIELD  NAME                    LENGTH CONTENTS
 1- 3    1      label identifier          3    HDR
 4       2      label number              1    1
 5-21    3      file identifier          17    (file name) $
22-27    4      set identification        6    (file type) $
28-31    5      file section number       4    (0001-0002-nnnn)
32-35    6      file sequence number      4    (0001-0002-nnnn)
36-39    7      generation number         4    (file generation) $
40-41    8      generation version number 2    (version number) $
42-47    9      creation date             6    (ANSI standard date) $
48-53   10      expiration date           6    (spaces) $
54      11      accessibility             1    (space)
55-60   12      block count               6    000000
61-73   13      system code              13    SINTRAN III x
74-80   14      (not used)                7    (spaces)

$ field 3:

- An apostrophe is used to mark the end of the string. This character is not a part of the name. The unused part of a field is filled with spaces.

$ field 4:

- Only the first four characters are used in this field. If it is shorter than four characters, an apostrophe is used to mark the end of the string.

$ field 7:

- Only alphanumeric characters. The field is left justified, and an apostrophe is used to mark the end of the string. The character code in this field identifies a backup generation of files.

$ field 8:

- This field contains numbers from 1 to 99. The characters are left justified, and one digit numbers will have an apostrophe in the right character position. This number identifies different versions of files with identical file names (fields 3 and 4), and each version must be treated as an individual file.

$ fields 9 and 10:

- Creation and expiration dates are not used, and the fields will contain spaces.

$ field 13:

- The system code is SINTRAN III x, where x is the version code, for example J.

Second file header label

      SECOND FILE HEADER LABEL
POSITION FIELD NAME               LENGTH CONTENTS
1- 3      1    label identifier     3    HDR
4         2 label number            1    2
5         3 record format           1    U
6-10      4 block length            5    (no. of characters)
11-15     5 record length           5    (spaces)
16-50     6 reserved                     (name of owner) $
            for operating systems  35    & MAX BYTE POINTER)
51-52     7 (not used)              2    (spaces)
53-80     8 (not used)              2    (spaces)

$ field 6:

- Up to 16 alphanumeric characters, starting from position 16, identifying the owner of this file. If the name is shorter than 16 characters, an apostrophe is used to mark the end of name.

- 32-41 contain an ASCII representation of the maximum byte pointer of the file.

End of file label

               END OF FILE LABEL
POSITION FIELD  NAME                 LENGTH CONTENTS
 1- 3     1     label identifier       3    EOF
 4        2     label number           1    1
 5-54     3-11  (same as HDR1)        50    (corresponds to HDR1)
55-60    12     block count            6    (number of blocks)
61-80    13-14  (not used)            20    (spaces)

End of volume label

               END OF VOLUME LABEL
POSITION FIELD  NAME                 LENGTH CONTENTS
1-3      1      label identifier      3     EOV
4-80     2-14   same as EOF1         77     (corresponds to EOF1)

User label

                      USER LABEL
POSITION FIELD  NAME                LENGTH CONTENTS                                  
1-3      1      label identifier      3    UHL
4        2      label number          1    1           
5-80     3      user option          76    (file information) $

Explanation of field 3

- $ This field differs from the ANSI label standard. The field contains binary information for the ND BACKUP-SYSTEM and SINTRAN III file system.

Field 3:

       POSITION                         CONTENTS
WITHIN FIELD WITHIN LABEL
 1-2          5-6                version number of this file (1-255)
 3-4          7-8                total number of versions (1-255)
 5-8          9-12               file system standard creation date
 9-12        13-16               not used
13-76        17-80               SINTRAN III file system object entry

Nonstandard 'Hole' label

        NONSTANDARD 'HOLE' LABEL
POSITION FIELD    NAME             LENGTH CONTENTS
1-3       1       label identifier   3    HOL
4         2       label number       1    E
5-80      3       user option       76    (information) $

Explanation of field 3:

Field 3:

            POSITION                     CONTENTS
WITHIN FIELD WITHIN LABEL
 1-72         5-76         THIS BLOCK IS NOT PART OF THE DATA!
                           CHARACTERS 77-80 CONTAIN A NUMBER.
73-76        77-80         (32-bit binary number stating the logical
                           block number of the following data block)

Modifications for device backup

When a complete device unit is copied to a volume on disk or tape, the copy of the device unit/subunit on the volume is a normal volume file. If the backup system command DEVICE-COPY is used, the system itself decides whether a device should contain a volume or directory. The device should contain a volume if it is a streamer or magnetic tape, or if the source and destination devices have different sizes, and the destination device is not an entered directory while the source is.

The ANSI labels of a destination volume are then generated as follows:

VOL1 - Volume header label

FIELD 3 - Volume serial number {volume name).

Each time a device is copied to a volume, a new volume is automatically created by the system. The volume name is not explicitely given by the user, but decided by the system to be BACKUP.

FIELD 7 - Owner identification (name of owner).

The owner will be SYSTEM.

HDR1 - First file header label

FIELD 3 - File identifier (file name).

The file name is the name of the source device, for example DISC-75MB-1

FIELD 4 - Set identification (file type).

The file type is the unit/subunit identification given as, for instance, 1 (unit 1) if there are no subunits, or 1-2 if subunit 2 is copied, or 2R or 2F1 if the source is a Phoenix disk with removable (R) and fixed (F) subunits (2F1 means fixed subunit 1 of unit 2).

FIELDS 6, 7 and 8 - Sequence, generation and version numbers. Since a new volume is created each time a device is copied, the copy will be the first file on the volume, and thus the sequence number is 0001. When the DEVICE-COPY function is COPY, the generation and version numbers are set to 1.

HDR2 - Second file header label

FIELD 6 - Res. for operating systems (length: 35 bytes).

- Bytes 1 - 16:

  ND-CPU-xxxxxxxxx where xxx... is the CPU number of the system, left-justified and terminated by an apostrophe.
  This may be regarded as the owner of the file.

- Bytes 17 - 32:

  Directory name of the source.

- Byte 33:

   Directory index of the source.

- Bytes 34 - 35:

   File  system flag word (DFLAG)  of the directory table entry  of the source.

UHL1 - User label

FIELD 3 - User option (length: 76 bytes) - Bytes 1 - 2:

 Binary generation (= version) number.

- Bytes 3 - 4:

 Total number of generations.

- Bytes 5 - 8:

 Copy date in the standard SINTRAN III file system format.

- Bytes 9 - 12:

 Start page number. When the section number of the HDR1
 label is > 1 (the device copy is split into several vol-
 umes), this is the start page of the current section. If
 the section number is 1, and the start page is >< to the
 file pointer of the device object entry (see below), then
 the source device was a main directory containing
 SINTRAN III, and:

   start page - file pointer =
     logical number of first page behind (SYSTEM)SEGFILO:DATA

Thus, when recovering this file, one may skip SINTRAN III.

- Bytes 13 - 76:

  Device object entry (length: 64 bytes). This object entry
  is constructed by the backup system, and is quite similar
  to an ordinary SINTRAN III file system object entry. It
  describes the source device, and contains the following:

Device object entry in field 3

POSITION   CONTENTS
IN OBJECT
ENTRY

 1- 2      Flag word (NFLAG) of the name table entry of the device
 3-18      Device name
19-22      Unit subunit: 1', 2-1', 1R', 2F1' (see HDR1 file type)
23-26      Zero (version chain pointers)
27-28      Access word is 1, i.e. only owner read access
29-30      File attribute word is 20B, i.e. contiguous file
31-32      Unit and device number (DUNIT) from the directory entry
33-34      Term.no. and user index of the user who made the copy
35-36      Subunit and name index (LUNIT) from the directory entry
37-52      Zero (open counts and dates)
53-56      Number of pages in file (source device)
57-60      Max. byte pointer = number of pages * 2048 - 1
61-64      File pointer. Start page of source device on disk

Special rules for floppy volumes

- Only floppy formats 0 and 17 are allowed.

- Each label group (HDR1 HDR2 UHL1) is written in one record (sector on the floppy).

- A HOLE label is written as a deleted record to distinguish it from data, and the EOF mark preceding the EOF1-1abel is written as a deleted record containing the identifier 'EOF*' in the first 4 bytes of the record, to distinguish if from a HOLE label.

Special rules for disk and streamer

A11 the labels in front of, between, or at the end of files on disk or streamer volumes are written in one sector, and they have fixed positions in the sector. An exception from this may be the case where one needs several UHLn labels (n>1) to contain a logical bit file (based on the index blocks) of large indexed files, when normal file backup on disk or streamer volumes is implemented.

BEGINNING OF VOLUME.

VOL1 Zero-word Zero-word HDR1 HDR2 UHL1 EOF-word

The Zero-word contains only 0, while the EOF-word contains binary 3 (EOF) in the first byte, and an asterisk (*) in the last byte.

BETWEEN FILES.

Eof-word EOF1 Eof-word HDR1 HDR2 UHL1 Eof-word

END OF VOLUME.

           EOF1
Eof-word  < or > Eof-word Eof-word            
           EOV1

if the volume is not empty. If the volume is empty, the contents of the first sector is:

VOL1 Zero-word Eof-word Eof-word

The labels and Zero/Eof-words in front of, or between files will thus take 4 * 80 bytes + 3 * 2 bytes = 326 bytes = 163 words of a sector.

Product number

• ND-10337
• ND-210337