Victor, First thing, both the tape management systems and SMF records get information such as "blocksize" from the application control blocks (DEB, DCB, JFCB for example). However, when EXCP or even SIO commands are issued by the application the actual block-size of the data being written can be more or less than what is being reported. In your example, DFDSS is saying 256k blocks will be written while in reality the size of the physical blocks of data actually written is about 72% of that. Likewise, when FDR says it is writing in 32k blocks it is often times writing data in 47k blocks on the tape. Why, because the application (DFDSS or FDR) open'ed the file by telling the operating system one thing and then did something else.
Now, the question that has never been answered is WHY do you need to know the true blocksize of each block of data written on the tape. Remember, using common I/O commands such as EXCP, I could easily write the data so that each block was a slightly different blocksize. One block would be 262144, one block would be 247137, one block would be 249152. What difference does it make? If you are attempting to copy the data, the copy program (Tivoli Tape Optimizer or CA Copycat for example) will attempt to read the largest supported block of data for that specific device type and the I/O returned would indicate exactly what the size of each block of data was. Now, the blocksize has NOTHING to do with question of compression done at the hardware level (COMP or NOCOMP). That is controlled by JCL parameters and system parameters. So, in the case of DFDSS it might write data in 200,00k blocks that are then compressed in the control unit and the compressed data is then written in physical 256k blocks of compressed data. So the physical block of 256k on the tape might actually control 5 blocks of compressed data where each un-compressed block of data is about 200,000k in size. So the operating system "block" written down the channel has nothing to do with the physical "block" of compressed data written on the tape. Russell Witt CA Technologies On 05/06/14, Victor Zhang<victor_wor...@aliyun.com> wrote: In RMM, I checked the block size = 262144, which is 256KB in size. And if we can't trust it, and does dss program record block size in SMF 21 record? Or Is there another method to check tape block size, ie using DITTO? Regards Victor ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
