How To Specify Variable Block Size for Tape Backup FunctionsLast reviewed: December 23, 1996Article ID: Q161338 |
The information in this article applies to:
SUMMARYIn Windows NT, tape backup API's can be configured to manipulate tapes that use either fixed block size or variable block size. The default mode for tape backup functions is for fixed block size. In this configuration, the size of all data buffers used for reading and writing must be a multiple of the block size. To determine the fixed block size, the GetTapeParameters function must be used. Specifying the GET_TAPE_MEDIA_INFORMATION operation will yield a TAPE_GET_MEDIA_PARAMETERS structure. The BlockSize member of this structure reports the block size of the tape. The size of buffers used in read and write operations must be a multiple of the block size. This mode allows multiple blocks to be transferred in a single operation. If, however, variable block size is desired, then more steps are involved to successfully manipulate a tape. First, the drive must be interrogated to determine if it is capable of supporting variable block size. Second, the block size must be set for variable size. This article discusses the details of these operations.
MORE INFORMATIONTo determine if a tape drive can support variable block size, the GetTapeParameters function must be called. Specifying the GET_TAPE_DRIVE_INFORMATION operation will yield a TAPE_GET_DRIVE_PARAMETERS structure. The dwFeaturesLow member of this structure is a flag that will have the TAPE_DRIVE_VARIABLE_BLOCK bit set if the drive can support variable block size. The following sample code shows how to use the function:
BOOL DriveSupportsVariableBlockSize() { TAPE_GET_DRIVE_PARAMETERS DriveParms; DWORD dwBufferSize; DWORD dwErrorCode; dwBufferSize = sizeof(DriveParms); dwErrorCode = GetTapeParameters( ghTape, GET_TAPE_DRIVE_INFORMATION, &dwBufferSize, &DriveParms ); if (dwErrorCode != NO_ERROR) { // An error occurred calling the function. // The error should be handled by some error handling function. return FALSE; } if (DriveParms.dwFeaturesLow & TAPE_DRIVE_VARIABLE_BLOCK) return TRUE; else return FALSE; }If the drive reports that it is not capable of supporting variable block size, then variable block size cannot be used. Tapes that use variable block size will not be usable with the tape drive. If, on the other hand, the tape drive indicates that it is capable of supporting variable block size, then the next step is to set the tape media parameters to variable block size. The SetTapeParameters function must be called specifying the SET_TAPE_MEDIA_INFORMATION operation. The function requires the use of a TAPE_SET_MEDIA_PARAMETERS structure. The BlockSize member of the structure must be set to the desired block size. A block size of 0 indicates variable block size. Any other value sets the media parameters to fixed block size. The following sample code demonstrates how to use the function:
BOOL SetTapeBlockSize(DWORD dwNewBlockSize) { TAPE_SET_MEDIA_PARAMETERS MediaParms; DWORD dwErrorCode; MediaParms.BlockSize = dwNewBlockSize; dwErrorCode = SetTapeParameters( ghTape, SET_TAPE_MEDIA_INFORMATION, &MediaParms ); if (dwErrorCode != NO_ERROR) { // An error occurred calling the function. // The error should be handled by some error handling function. return FALSE; } return TRUE; }When writing to a tape using variable block size, each write operation creates a new data block on the tape. The size of the block is the size of the write operation. There is an exception to this when a single write may create more than one block. For additional information on this exception, please see the following article in the Microsoft Knowledge Base:
ARTICLE-ID: Q152518 TITLE : PRB: Tape Variable Blocksize Limited to 64K BytesReading a tape containing variable sized blocks can be accomplished even without knowing what size the blocks are. If a buffer is large enough to read the data in a block, then the data will be read without any errors. If the buffer is larger than a block, then only data in a single block will be read and the tape will be advanced to the next block. The size of the block is returned by the read operation in the &dwBytesRead parameter. If, on the other hand, a data buffer is too small to contain all of the data in a block, then a couple of things occur. First, the data buffer will contain data from the tape, but the read operation will fail and GetLastError will return ERROR_MORE_DATA. This error value indicates that there is more data in the block to be read. Second, the tape is advanced to the next block. To re-read the previous block, the tape must be repositioned to the desired block and a larger buffer must be specified. It is best to specify as large a buffer as possible so that this does not occur. If a tape has fixed size blocks, but the tape media parameters are set to variable block size, then no assumptions are made regarding the size of the blocks on the tape. Each read operation behaves as described above. The size of the blocks on the tape are treated as variable, but happen to be the same size. If a tape has variable size blocks, but the tape media parameters are set to fixed block size, then the size of all blocks on the tape are expected to be the same fixed size. Reading a block of a tape in this situation will fail and GetLastError will return ERROR_INVALID_BLOCK_LENGTH. The only exception to this is if the block size in the media parameters is the same as the size of the variable block and the size of the read buffer happens to be a multiple of the size of the variable block.
|
KBCategory: kbprg kbhowto
© 1998 Microsoft Corporation. All rights reserved. Terms of Use. |