AnsweredAssumed Answered

After HTTP call retry not all documents are present

Question asked by martijn.keizer147426 on Nov 17, 2016
Latest reply on Nov 22, 2016 by James Ahlborn

The process I have is as following. At first a list of items to download is retrieved. This list is then split into individual items, and those then get retrieved by a HTTP call. I have a retry mechanism in place that looks as such:

In essence, from what I understand all items are retrived. The downloads that succeed get checked for content type (it should be application/octet-stream, anything else is considered a failure and should be retried) and then written to the document cache. Any retries are later added to the same cache, and when the downloads have all succeeded, this cache is emptied and all downloads are returned (this branch is not shown).

 

When simulating a failure in the HTTP call by turning the download service off for a short while during process execution it doesn't seem to write the initial succeeded cases to the document cache, as shown by the process state tree below.

I can't seem to be able to figure out what happens to the decision step when the 19 documents of the first try reach it. Anyone knows what's going on? Or how to circumvent this issue?

Outcomes