On Tue, Oct 07, 2008 at 04:56:12PM -0400, Johnson, Alex P (IPS) wrote: > My two cents. > > The application that creates the file should move it and signal that the > move is complete. On many projects, I've used ftp to move the file and > rsh to run the application that processes the file. This combination > ensures that the file is fully transferred before processing begins. > > Failing that, I will monitor the 'modification date' on the transferred > file and wait some period after that date before processing the file. > > I agree with Alex on his points. But in practice I have run into a scenario where we did not have control of the application on the Windows side. So, what I did was have the DCS side transfer the file, and the Windows side delete the file after it had processed it. At least I could check (using ftp) for the existence of the file, and not transfer a new on if the Windows app has died. I admit it's not an elegant solution, since I have no control over the Windows side trying to "consume" an incomplete file, since ftp is not atomic. My original design had their side checking for a lock file, that I would have transfered _after_ the data file, and them then deleting this lock file when they were done consuming the data, but it did not get implemented that way, unfortunately. -- One of the main causes of the fall of the roman empire was that, lacking zero, they had no way to indicate successful termination of their C programs. _______________________________________________________________________ This mailing list is neither sponsored nor endorsed by Invensys Process Systems (formerly The Foxboro Company). Use the info you obtain here at your own risks. Read http://www.thecassandraproject.org/disclaimer.html foxboro mailing list: //www.freelists.org/list/foxboro to subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join to unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave