I have two WCF services one public facing which proxies to a 'private' service which actually does the real work. The services exposed have identical interfaces and (practically) identical configs.
One of the interfaces is a streaming service using MTOM because there is a method to upload rather large binary files (and some smaller files too) the method is "AddAsset" it takes a binary array and a name and file type.
The client application receives a Communication Exception when attempting to call the AddAsset method but not always: If the client is attempting to upload, say, 20 images it will open the connection to the host and send one image at a time. Most images
succeed BUT if an exception is received we have written a "retry" loop which gives it three (set by local config) more attempts. Here are some facts we have discovered:
- If a particular image fails once it will always fail in the client application it was run from
- The failures only occur from client applications based in one site - all other sites we have tried work no matter the internet connection speed
- tracing at the proxy demonstrates the exception
- tracing at the backend service does not log the exception - in fact does not log enough I think?
- Successive "retries" for an image always fail - that image will
View Complete Post