AIIM-MS66-1999.pdf

STD-AIIM MSbb-ENGL 1997 11111 1012348 O501802 978 1111 ANSIIAIIM MS66-1999 American National Standard for Information and Image Management - Metadata for Interchange of Files on Sequential Storage Media Between File Storage Management Systems (FSMSs) An American National Standard prepared by the Association for Information and Image Management International Approved as an American National Standard April 6, 1999 Abstract: This document describes a standard for specifying metadata that describes how a File Storage Management System (FSMS) has stored files on sequential media. This metadata description is independent of any particular proprietary metadata format, but it may be used to describe the proprietary format that an FSMS uses to write files to sequential media. COPYRIGHT Association for Information an identifier that is unique among all objects in the universe of discourse. In this con- text, a UID is unique among all of the objects in the metadata export. It need not be reproducible, since one acceptable method of generating such an I D involves using a time stamp. 4.34 user with respect to the metadata export, the person who causes the metadata export to be written and in so doing makes the choices of record types. More generally, a person or program that commands an appli- cation. 5 Abbreviations and acronyms 5.1 DCE Distributed Computing Environment. 8 ASSOCIATION FOR INFORMATION AND IMAGE MANAGEMENT INTERNATIONAL COPYRIGHT Association for Information however, in the representations of clauses 7 through IO, they are all upper case. This standard does not impose a limit to the length of a record or to the length of any of the information fields. The only length restrictions are imposed by the operating system that writes the metadata. Simi- larly, the size of the numbers that are likely to be encountered in the metadata is not limited by the stan- ASSOCIATION FOR INFORMATION AND IMAGE MANAGEMENT INTERNATIONAL 9 COPYRIGHT Association for Information 2. metadata describing the removable media being exchanged; 3. metadata describing the file segments on the removable media being exchanged; 4. metadata describing the directory and file structure being exchanged. The metadata export shall consist of four sections, one corresponding to each of the four types. If the metadata export is presented as one file, then the first section shall contain all of the metadata of the first type, and each of the other sections shall contain all of the metadata of one of the other types and may occur in any order. For metadata export, it is also acceptable that each of these sections be in a separate file if the files are appropriately identified. 6.5 Order If the order of a particular record is not specified, it may appear in any order, if that does not conflict with records whose order is specified. Similarly, if the order of fields is not specified, then they may be in any order. For example, in the second section of the metadata, all records following a Cartridge Identifier record refer to the cartridge identified in it, until another Cartridge Identifier record is encountered. The records between the Cartridge Identifier records may be in any order. 6.6 Counting and counts All counting in the metadata shall be zero-based. That is, the first element shall be element O; the second element shall be element 1, etc. For example, a stripe set consisting of five cartridges shall have the 10 ASSOCIATION FOR INFORMATION AND IMAGE MANAGEMENT INTERNATIONAL COPYRIGHT Association for Information NO if not. Record category: Required. Example: CART-DRlVE-TYPE0MFR0ABC0MODEL0XYZ40000 TY PE0MNZ0DRIVE-COMPRESSlON0YES0§ Rationale: This record, with the record described in 8.2, is designed to remove any ambiguity in speci- fying a drive for reading or writing the cartridge. 8.4 Cartridge layout Record name: CART-LAYOUT Description: This record describes the high level layout of the data on the cartridge. For the purposes of this standard, all cartridges are assumed to be partitioned into at least one partition, even if partitioning is not supported by that cartridge technology. Named fields: NO-PARTITIONS: The number of partitions on the cartridge. This number shall always be at least 1. NO-SIDES: The number of sides on the cartridge. This number shall always be at least I. SIDE-NO, PARTITION-NO, PARTITION -NAME, LABEL-TYPE, INTERNAL-LABEL, SIZE-MB, FREE-SPACE-MB: These fields shall appear with the SIDE-NO and PARTITION-NO fields first in that order. The remaining fields (LABEL-TYPE, INTERNAL-LABEL, SIZE-MB, FREE-SPACE-MB) may appear in any order, and all refer to the partition on the side specified by the two previous fields. This entire group of fields shall be repeated for as many partitions as there are on the cartridge. SIDE-NO: The side number. PARTITION-NO: The number of one of the partitions on the tape. Partitions are to be numbered con- secutively, in physical order, O, 1, 2, 3, . n-I, where there are n partitions on a side of the cartridge. PARTITION-NAME: The name, if any, of the partition. If the partition is not named, then this field may be omitted. SIZE-MB: The estimated capacity of the partition, in MB, represented as an integer, rounded to the nearest MB (1 024 x 1024). FREE-SPACE-MB: The estimated free space of the partition, in MB, represented as an integer, rounded to the nearest MB (1 024 x 1024). LABEL-TYPE: The name of the label format, assuming that there is an internal label on the cartridge or partition. If this label conforms to a standard, this is the name and number of the standard. Otherwise the label format shall be named and described in a separate document, and the name specified in that document shall be used for this field. INTERNAL-LABEL: The contents of the internal label, sufficient to do label checking. ASSOCIATION FOR INFORMATION AND IMAGE MANAGEMENT INTERNATIONAL 15 COPYRIGHT Association for Information oth- erwise, of course, there is only one set; that is, the CART-ID and the remaining fields of this record de- scribe the starting location of the file segment on the specified cartridge. For a striped file segment, the cartridges and the starting addresses shall be specified in the order in which data is written to the car- tridges. All of the following fields refer to the cartridge identified in this field. Note also that if the file segment has a header part and/or a trailer part, the header and trailer shall be specified in the header and trailer records; that the starting address specified in this record is the first byte of the header; but that the length specified above does not include the length in bytes of any header or trailer blocks. CART-ID: The ID of the cartridge containing this file segment. The value of this field shall match some entry in the Cartridge identifier record specified in 8.1. PARTITION: This specifies the partition number on the cartridge in which the file segment starts. If this field is not specified, it is assumed that the file segment begins in partition zero. SIDE: This specifies the side number on the cartridge in which the file segment starts. If this field is not specified, it is assumed that the file segment begins in side zero. TAPE-MARK: This specifies the number of tape marks within the partition in which the file segment starts to skip in positioning to the file. If this field is not specified, it is assumed that the file segment be- gins in file zero. BLOCK: This specifies the block after the tape mark, within the specified partition, in which the file seg- ment begins. If this field is not specified, it is assumed that the file segment begins in block zero. In gen- eral, a block is created by physical I/O to the sequential media by the device driver. RECORD: This specifies the record within the block in which the file segment begins. If this field is not specified, it is assumed that the file segment begins in record zero. For the purposes of this standard, a record is a part of a block. Note that not all sequential media technologies preserve record boundaries. BYTE: This specifies the byte within the block or record in which the file segment begins. If this field is not specified, it is assumed that the file segment begins in byte zero. COOKIE: Application-dependent value for fast positioning on a cartridge. The exact nature of this entry shall be documented separately by the exporting FSMS. Not all cartridges support this. Record category: Required. Examples: - For non-striped file segments on a cartridge having cartridge identifier AB0003. The file segment begins after the first tape mark in the third partition, begins immediately at the begin- ning of the partition, is of length 6083 bytes, and represents bytes 2365 through 8447 of the original file: FS-I NFO0OFFSET023640LENGTH060830CART-I D0AB00030PARTITI ON020s The file segment is the first thing on the cartridge, is of length 2089, and represents bytes 10365 through 12453 of the original file: 24 ASSOCIATION FOR INFORMATION AND IMAGE MANAGEMENT INTERNATIONAL COPYRIGHT Association for Information & Imaging Management Licensed by Information Handling Services COPYRIGHT Association for Information & Imaging Management Licensed by Information Handling Services STDOAIIM MSbb-ENGL 1797 3032346 0503834 488 0 ANSIIAIIM MS66-1999 - METADATA FOR INTERCHANGE OF FILES ON SEQUENTIAL STORAGE MEDIA BETWEEN FILE STORAGE MANAGEMENT SYSTEMS (FSMSs) FS-I NFO0OFFSET0 103640LENGTH020890CART-I D0AB000309 The file segment is on a non-partitioned cartridge, is after the 13th tape mark, is of length 1024, and rep- resents bytes 89455 through 90478 of the original file: FSINFO0OFFSET0894540LENGTH010240CARTID0AB00030TAPEMARK0 1 309 The file segment is on the fifth partition, begins at the sixth block, is of length 5349, and represents bytes 2170 through 751 9: FSINFO0OFFSET021690LENGTH053490CARTID0AB00030PARTITION040 BLOCK0509 - For a striped file segment on cartridges having cartridge identifiers AB0010, ABOOII, AB0012, and where the file segment is striped across those cartridges in that order. The file segment is after the third tape mark on each of the cartridges, it is the beginning of the file, the file segment length is 2397960, and stripe block length is 159864 bytes: FS-INFO0STRIPE-WIDTH030STRIPE-BLOCK-LENGTH01598640OFFSET0 00LENGTH023979600CART-I D0AB001 O0TAPE-MARK030CART-I D0AB0011 0TAPE-MARK030CART-ID0ABOOl20TAPE-MARK030§ Rationale: This record allows for the specification of a file segment address on a cartridge in a number of different ways. Different technologies allow for different addressing or for a variety of addressing schemes. At least one helical scan implementation, for instance, maintains the original record bounda- ries as written by the host system. 9.6 File segment header Record name: FS-HEADER Description: This record specifies the length in bytes of the header, if any, which is added by the FSMS. The number of bytes in the header is not included in the specification of the file segment length. Named fields: None. Record category: Required, if there is a header. Examples: If there is a header for the file segment of length 12 bytes, FS-HEADER01205 Rationale: Some implementations add a header containing control information to each file segment. This is not to be confused with implementations that place headers on data blocks, which may not be file segments. ASSOCIATION FOR INFORMATION AND IMAGE MANAGEMENT INTERNATIONAL 25 COPYRIGHT Association for Information & Imaging Management Licensed by Information Handling Services COPYRIGHT Association for Information & Imaging Management Licensed by Information Handling Services ANSI/AIIM MS66-1999 - METADATA FOR INTERCHANGE OF FILES ON SEQUENTIAL STORAGE MEDIA BETWEEN FILE STORAGE MANAGEMENTSYSTEMS (FSMSs) 9.7 File segment trailer Record name: FS-TRAILER Description: This record specifies the length in bytes of the trailer, if any, which is added by the FSMS. The number of bytes in the trailer is not included in the specification of the file segment length. Named fields: None. Record category: Required, if there is a trailer. Examples: If there is a trailer of length 12 bytes, FS-TRAILER0120fj Rationale: Some implementations add a trailer containing control information to each file segment. This is not to be confused with implementations that place trailers on data blocks, which may not be file seg- ments. 10 Metadata describing the directory and file structure being exchanged There are only two record types defined in this clause. The first describes an object in the file system that may be a directory, an ordinary file, or some kind of special file. The second record type applies only to objects such as directories or tar files that are known by the FSMS to contain files. It gives a list of files contained in the object described in the previous record. Other than this, there is no restriction on the order of the records in this clause. The names of all records specified in this clause begin with the characters “D-“ 10.1 Object attribute list Record name: D-OBJ-ATTR-LIST Description: This record identifies an object and gives information about it. Hard links are identified by two or more files having the same UID but different fully qualified path names. The absence of clauses describing Access Control Lists, Discretionary Access Control, Mandatory Ac- cess Control, and other security information is deliberate because there is no common practice evident among different FSMS or operating system vendors. This type of information should be transferred ex- plicitly and under the direction of the appropriate security administrators. It would not be prudent to ex- pect an importing FSMS to update or change security information in any way. Named fields: UID: An identifier of the object (file, directory, etc.) which is unique within the objects described by this metadata. NAME: The fully qualified name of this object. OBJ-TYPE: One of the following: DIR for a directory, FILE for an ordinary file, SYM for a symbolic link, and CONTAINER for a file that the FSMS uses to hold other files. 26 ASSOCIATION FOR INFORMATION AND IMAGE MANAGEMENT INTERNATIONAL COPYRIGHT Association for Information & Imaging Management Licensed by Information Handling Services COPYRIGHT Association for Information & Imaging Management Licensed by Information Handling Services STD-AI11 MSbb-ENGL 1777 . I I LO12348 050Lô3b 250 E ANSIIAIIM MS66-1999 - METADATA FOR INTERCHANGE OF FILES ON SEQUENTIALSTORAGE MEDIA BETWEEN FILE STORAGE MANAGEMENTSYSTEMS (FSMSc) PERMISSIONS: The octal representation (in ASCII or Unicode characters) of the absolute mode of the POSIX access permissions on this object. Deviations from what is below shall be documented sepa- rately. An absolute mode is specified using octal numbers nnnn where n is a number from O to 7. An absolute mode is constructed from the OR of any of the following modes: 4000: Set user ID on execution. 2000: Set group ID on execution. 040