IBM Mainframe Forum

Every month, we FTP over 200 files from the mainframe to a Windows server and it can take from 8-12 hours. Many of the files are very large and have to be split for the FTP. We are able to compare record counts (expected and received) and a few of the files have incorrect record counts and must be re-FTP'd. It's usually less than 50 records per file and happens to a random 2-10 files.

We tried limiting the number of FTP stream from 5 to 3 but that had no effect.

Where would be the first places to research to identify the cause? Reftp is a time consuming process because of the numerous large files.

MK

Is using PKZIP/GZIP or something similar to greatly reduce the filesizes not an option?

Is it possible to verify the result of FTP for each particular file?
Either on the fly, just after every FTP command, or after the fact - by the FTP log analyzis?

It should show the files in error, to repeat this FTP, either manually, or automatically.
Even for thousands of FTP...