3

At my company we have some vendors that we transfer data to and from. Sometimes the data is pulled into our local SQL database for business reporting. Other times we pull data from one vendor, transform it, and then transfer it to another vendor's FTP server.

The guy I replaced has a couple of generic SFTP Push/Pull console applications that transfer data to/from these vendors. Then he has other applications that do the data imports to the SQL database or transforms the data and leaves it out in a directory to be pushed to the vendor.

Every once in a while we will have problems with these processes not finding the file needed and I have to go back and run these manually to load the data. It seems to me that it would be more reliable if these processes just did their own FTP push/pull functions so we don't run into scheduling problems. Are there any standard practices that I can implement that would be more reliable or do I need to just tweak what I have now? I'm in a Windows/.NET environment by the way.

  • 2
    Your question needs to be more specific than this. Solutions like this tend to be brittle, because they are often cobbled together with ill-fitting, off-the-shelf tools run in batch processes. The best practice is to write a real application that is specifically designed to do what you want it to do. – Robert Harvey Nov 19 '13 at 17:31
  • Are you monitoring the folders for the presence of these files or are you just triggering everything with some sort of timer? The .NET framework has plenty of tools to do it either way. – JeffO Nov 19 '13 at 18:08
  • @JeffO, They are all just triggered by scheduled jobs hoping for a file to be there. I was thinking of just including the ftp logic within each dependent program or doing something to listen to the folders and then start my data load process. Before I go off and start experimenting I want to see if there are any standard practices I can implement. – programmer Nov 19 '13 at 18:38
  • 1
    This depends a lot on the vendors involved and agreements you have with them. An ideal solution would be a kind of web service (or alternatively something like EDI) where you have a more direct access and data is actually generated whenever you request it, like maybe get all new orders since last request. – thorsten müller Nov 19 '13 at 19:02
3

First of all, having the ftp push/pull mechanism separated from the core processing sounds to me a good design, since it will allow to test the core processing separately and plug the parts easily together in a different way if needed. This is a good example of separation of concerns.

Every once in a while we will have problems with these processes not finding the file needed

Before thinking about any solution with the chance of causing probably more problems than it will solve, make sure you know what the root cause of the problem is. Is it because job A (pulling the data) puts the file in a wrong folder where job B (pushing the file data) does not expect it? Then you need a better way to pass the file path from job A to job B in a reliable way.

Or is it because sometimes job B starts too early, before the output of job A has arrived completely? Well, then you need a better mechanism to trigger the start of job B. Is it not possible to put A and B in a command script which makes sure B does only start when A is complete? Maybe you have to implement a polling mechanism in job B which makes sure it does not start its processing until the output of job A is available. Maybe you have to implement a loop around job A to make sure it will try to download a file again when the first attempt has failed. It may be a good idea to let the ftp process download all data into a temporary file first, and rename that file as a final step when it is complete. Renaming is an atomic operation on most file systems, so this makes the file only visible to the following processes when it is ready for further processing. Another possible technique is to work with some "lock files", prohibiting shared access to a file "X" as long as "X.lock" exists.

So, IMHO the architecture you described is not brittle per se, but you have to provide a reasonable amount of synchronization and failure tolerance around your processes.

  • Thanks doc, my problem was "...sometimes job B starts too early..." In one process I did add a polling mechanism to check if the file is there before it starts so I'll probably add polling to the rest also. I'll also update the ftp process to download the files to temporary files and rename them upon completion. – programmer Nov 19 '13 at 22:50
  • 1
    @JasonHolland: sounds like you could make that part generic - write a tool or script which gets a file name (or a list of file names) and the name/parameters of a command line program, and it starts the latter as soon as the file (or files) are available. Keeping the individual steps of your process chain independent follows the idea of the well known "pipes-and-filters" pattern (see, for example, here: eaipatterns.com/PipesAndFilters.html). – Doc Brown Nov 20 '13 at 6:41
  • Thanks for the additional advice and link to the pipes-and-filters pattern. I know I need to fix what I have now but I wanted to see if I could find some standards or patterns to apply to this problem rather than go off half-cocked solving it. – programmer Nov 20 '13 at 14:18
  • @JasonHolland It seems that some research into multi-threading may provide you with a standard reference for certain concepts that apply to this; locking; barriers; avoiding deadlocks. – Andyz Smith Nov 20 '13 at 15:21
  • There's a nice tool in the SQL server ecosystem to help with this: en.wikipedia.org/wiki/SQL_Server_Integration_Services – Daniel J. Pritchett Nov 20 '13 at 19:07
1

Are there any standard practices that I can implement that would be more reliable or do I need to just tweak what I have now?

Tweak what you have now. For this kind of thing you probably have quite a few custom requirements - retries, availability windows, notifications, backing up, zipping. If it's working 90% of the time see if you can get it up to 99% of the time. Add a whole bunch of logging and exception handling and take it from there. Maybe run it manually yourself every day for 2 weeks to see if you can get it to fail.

I have yet to find a great application that takes care of SFTP'ing stuff to and from other servers with the option of setting availability windows, retries, notifications, etc. I believe SQL Server can do it. AFAIK, there's no "standard". Hmm maybe I should code something up and sell it :)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.