Activity timeout on admin console 1.8rc2

I’ve noticed that the timeout value while in the admin console is painfully short, sometimes after only 15 seconds of inactivity. Sometimes when I log back in, I get "Unexpected Error, please contact support. "
The log file says:
2897683 906f809 2017-01-30 13:47 WEB GET ERROR /region/preview/71 Template Missing Exception Type: Xibo\Exception\ControllerNotImplemented
2897682 5a8b99e 2017-01-30 13:47 WEB GET ERROR /library/download/35 Image source not readable Exception Type: Intervention\Image\Exception\NotReadableException
2897681 291bceb 2017-01-30 13:45 WEB GET ERROR /library/download/138 Error reading session: Different UserAgent

This is a layout I created Friday and am now trying to update with fresh material. timeout seems to happen from a remote PC and not on the server console.

The activity timeout is determined by your PHP configuration I believe.

You need to ensure your PHP session length is set to whatever timeout you want.

I’ve determined it’s not related to a PHP timeout, but if I try to edit any layout remotely it kicks me out to the login screen and shows errors in the log:

This was from just opening the layout on my screen and within seconds I was back to the login screen.

Your browser is sending a different user agent with each request - that is preventing your session from being continued on the grounds that it could a security breach.

What browser are you using please?

IE 11 on windows 10. Chrome won’t let me log in at all.

You need to track down why your user agent is changing with each request. It could be related to a browser extension, or any intermediate proxy server you’re using.

Chrome is definitely supported so it suggests there is some issue in your setup.

edit: I found a solution of sorts. troubleshooting:

I tried IE in safe mode, different operating systems (w10 x64, w7 x32) and from different subnets on the network. We don’t have an internal proxy server.
On my test server (fresh DB on same Xampp setup) it worked when using the IP.
On my upgrade to 1.8RC2, it fails using the server name.
It works fine in 1.79 on same server.

In wireshark I see the user-agent switch from
5670 Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko
5695 Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 10.0; WOW64; Trident/7.0; .NET4.0C; .NET4.0E; .NET CLR 2.0.50727; .NET CLR 3.0.30729; .NET CLR 3.5.30729)

It shouldn’t change. I found this article dealing with User-agent strings:

It mentions Compatibility mode, which shouldn’t be set on the devices I’m using. a check of the settings showed no sites in the list, but a checkbox 'Display Intranet sites in Compatibility View" was checked.

Unchecking that box has so far made it work in IE 11.

I will test more, but I think that fixed it.

additional info:
I’m using FQDN in the URL. It seems that something in the layout or php software is changing the URL so that the browser thinks it’s no longer in the intranet. The IP in the URL always uses the Internet compatibility mode settings. The cookie is the same throughout the transactions.

You definitely don’t want compatibility view.

We don’t care what the browser user agent is, but we do care if it changes. Nothing in Xibo can persuade your browser to send a different user agent. That’s completely up to the browser.

What I suspect is happening is you’re moving between Intranet zone and Internet zone due to the way you have your CMS or DNS configured, and so it’s seeing the swap.

Make sure you’re always using FQDN, and that the browser always sees that as Internet zone, and it should be fine.