Thanks. If I get a chance I'll run up a fresh copy of LS3 in a dev environment and test it to try and get a handle on what it does.

Just thought seen as I was here I'd post an update on this. It appears this problem occurs if someone starts a survey in one browser, completes it in another browser and then goes back to the first browser and tries to complete it once again.

Odd behaviour I know.

The problem is exacerbated by a our load balancer causing two session files being created because each browser gets a different session due to being unrecognised.

However that should not make a difference once the survey has been closed the second session should be invalidated when the user attempts to load the page. As it is LS tries to put the data into the DB and its only then that it fails and sends a message to the administrator instead of simply advising the user that this survey has already been done.

egisupport wrote: Just thought seen as I was here I'd post an update on this. It appears this problem occurs if someone starts a survey in one browser, completes it in another browser and then goes back to the first browser and tries to complete it once again.

Yes : multiple access is always a pain …


It's the reason why relmoadAnyResponse plugin add a control before reload a response (by token too)
gitlab.com/SondagesPro/coreAndTools/relo...AnyResponse.php#L935

In such case : if browser #1 are not closed, user receive a 409 header when try to open on a second browser.

With a delay.