I guess I don't understand. We usually have mainframe jobs that pick up files from our ftp server after Mercator process them. I'm not a mainframer, so I never FTP'd directly to the MF. What do you mean MF processing handling the layout? JCL?
After more thought, as long as the dataset is allocated properly (size, name) etc, you can simple move a file with the same qualifiers (ex. edi.mydata.tothemf) and it will automatically FTP it in that pre-alocates dataset?
Ok, here is a simple put and get example (command line FTP on Windows)
ftp
open your_mf_host_name
user johndoe
password whatever
put c:\test.txt test.text
Now - if the target file (on the mainframe) does not exist then you will need to 'add' allocation attributes via a QUOTE SITE command before executing the put command.
Another example
quote site lrecl=80 recfm=fb
put c:\test.txt test.text
Also - depending upon how your mainframe FTP daemon is setup - Default allocation attributes etc.. you may need to do some experimenting with some puts and gets and also talk to the resident 'systems programmer/staff' to get the default attributes. These examples will allocate (create) a dataset on the mainframe with my HLQ (High Level Qualifier) - usually your TSO USERID that you log onto.
If you need to output (put) to a specific HLQ... then you need to wrap the target dataset in quotes..
Example;
put c:\test.txt 'TEST.TEXT.DATASET'
get is bascially the same man..
Here is IBM's URL to the Communications Server Bookshelf
In either case VSAM = no. The file must be defined before FTPing.
In the case of VSE = No. Again, file whether 6.7 is pushing or MF is pulling the file must be defined before the FTP will be successfull.
MVS, non-vsam = ? don't know. You might be able to allocate create the file and allocate on the fly. Don't know enough to answer your question.
Also, would you be delete/defining the file with each FTP? or, appending FTP data throughout the day until a batch is run on the MF? You'd need to define the file as 'reuse' when defining the file.
We've found it best to let the JCL define the files and perform a get or at then end of a batch cycle define a new file for the next day.
Finally, we found it easier or less envolved to let the JCL perform a get. For the map to push the files we need to have our MF admin perform additional setup related to the file and DD name.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.