Question

I am using connect direct with scp and trying to send some pdf files from unix to mainframes.

On unix end, I have archive containing pdfs which I am simply renaming to ABC.XYZ.LMN.PQR (mainframe file name) and then sending to mainframe.

The archive contains variable length pdf files.

However, the requirement is:

For any variable length file mainframe needs to know the maximum possible length, of any record in the file. For e.g. say the LRECL is 1950.

How to include LRECL as well when the pdf files inside the archive file to be sent is of variable length ?

Was it helpful?

Solution

An alternative would be to transfer the files to a Unix System Services file (z/OS Unix) instead of a "traditional" z/OS dataset. Then the folks on the mainframe side could use their utilities to copy the file to a "traditional" mainframe dataset if that's what they need.

OTHER TIPS

For Variable Blocked datasets Only! If your maximum record size is 1950 you will want to specify RECFM=VB,LRECL=1954 adding 4 bytes more than your maximum record. This 4 byte allowance is for the Record Descriptor Word (RDW). If you need to specify BLKSIZE then the minimum is the size of LRECL plus another 4 bytes.

So in your example, your JCL will have DCB parameters that looks like: RECFM=VB,LRECL=1954,BLKSIZE=1958

Ideally, for optimal storage, BLKSIZE should be set to the largest size that does not exceed the device specific recommendation. i.e. TAPE devices typically try to use BLKSIZE=32670 (32 * 1024K - 8 for RDW & BDW). Disk drives may vary but in our shop BLKSIZE=23476 is considered optimal.

Incorrect blocking factors can waste tremendous amounts of space. When in doubt, let your system defaults apply or consult with your local system guru's for their device specific recommendations.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top