when ever i turn on web protection for a rule users who can use internet through this rule can use whatsapp application on there phones or web whatsapp
i tried to make a workaround for web whatsapp and created a top rule that allow access to web whatsapp and turned off web protection and that solved web whatsapp problem
now my problem is with the application it self it wont work until i turn off the web protection
although i made exception for it in the PROTECT>Web>Exceptions and checked the log viewer and it is all green and all http and https scan & Decrypt are turned off
is there any solution for this issue ?
create a web exception with this urls:
Here the image. In my case works. I use decrypt and scan on my XG.
UTM Certified Architect - XG Certified Architect
i solved whatsapp application issue yesterday
i dont use "Any" as service
the problem solved when i added whatsapp application ports and both of them were working till this morning
but couple hours ago the QR code came to the surface again
any idea why this strange behavior from the firewall??
Sounds similar but not identical.
One thing I had in mind: It always had been Android devices with that problem, no one used an iPhone...
Any ideas? I am running out of ideas, logs dont show anything on this...
I could log while the error/delay occured.
Problem occured two times, please see below. There are NO red lines, but some entries right around the time the delay occurred. I am not sure what they tell me, or how to solve them. Logfile is clean as long as whatsapp runs smoothly. No idea why this temporarily works sometimes, sometimes not.
If I deactivate my Web Policy error NEVER occurs. My policy at the moment only has a default "allow all" and a deny "url1, url2, etc.".
I dont have any clue...
Thank you in advance for any kind of help!
No red lines means there was not a deliberate block, it does not mean there were not errors.
See the lines that have status_code="502". That's an error code. Its either an error code that is generated by the WhatsApp server and being passed via the proxy, or it is one that the proxy itself is generating because it has a problem with the connection.
The dst_ip looks fine and I suspect that pharming protection (the original thread, and something that is fixed) is not related.
There are two courses of right now. The first is to contact support and have them take a deeper look at your system, get debug level logs, etc. That can better determine if there is a config issue, a code issue, or just a straight out incompatibility.
The other is just say whatever and bypass the proxy for this type of connection. Which you might need to do anyway based on the investigation.
See https://community.sophos.com/kb/en-us/128173 section "Create a firewall rule for a website".
thanks a lot, Michael!
Will try both ways you suggested. If both wont work I think I will contact support...
Keeping you updated! Thanks a lot for assistance so far.
i have contacted them yesterday and they replied to check log viewer and check if there is anything blocked by web protection filter
and i have replied with there is not anything blocked and clarified all steps i took in attempt to solve the issue but still waiting for there answer
please dont forget to feedback us with there reply
sorry for bringing this thread up again
i was able to solve the problem one more time by add each category i want to block in separate rule inside the web policy
once i did it the whatsapp application worked smoothly without delay
try to do the same and let us know
Interesting, I had the same issue withe Whatsapp I disabled the pharming protection created exceptions and allowed the ports it worked for a while then the problem reappeared as you all said the log viewer doesn't help at all with this problem it is not showing any problem or error
I will try your method and let you know
try to do the same and let us know what happened
so I am not really making progress, thinking about opening a support case...
If have tried everything I could imaginge but still could not make any progress. error still occurs as soon as I activate web protection.
funnywhise those devices have a different error as soon as I deactivate the web policy, but that seems to be a different story.
I really dont have a this complicated setup, i am really getting frustrated with that xg. and obviously some of the solutions are working for some of us, but not for others... really confusing and log files do not help at all.
anyone else making progress with that whatsapp piece of sh*t???
It is the WhatsApp application doing HTTP requests to a WhatsApp server. It may not be following normal HTTP standards.
Can you try something:
Go into the console (not ssh shell). So in the menu choose (4 Device Console).
show httpset http add_via_header offset http relay_invalid_http_traffic on
Try again. If it is still broken, please revert the changes.
Next thing would be to see if the problem is with other ports. Are the ports listed here allowed through the firewall?
Michael, thank you in advance! Will try that step by step.
How ist the console setting to be assessed in terms of security? I mean, filtering out invalid http traffic, if you dont, is this weaking security somehow?
Thanks a lot Michael
Getting back once I have testet both suggestions and found something new. Great forum support, thank you!
It can potentially decrease security because that traffic cannot be scanned. Basically traffic that is using port 80/443 must conform to the HTTP standard for us to be able to scan the traffic. This setting is basically if there is non-HTTP traffic or malformed traffic does it get blocked or allowed.
An example is a type of media streaming called icycast, not commonly used these days.
The HTTP standard is the client sends GET http://somesite and the server responds with 200 OK.
With an icycast, the client sends GET http://somesite and the server responds with 200 ICY.
That response does not conform to the HTTP standard and is therefore not allowed. Turning this global option on would allow it.
so just to be sure, in order to undo the changes i have to:
show httpset http add_via_header onset http relay_invalid_http_traffic off
Correct, that will set them back to their defaults. Though it might now have been clear, when I wrote the original
that was so you could see and remember the currently set options before fiddling so you could return them back.
set http add_via_header offset http relay_invalid_http_traffic on
it did not work, unfortunately, this morning i had the same circumstances.
dumb question: if i allow services "all" in firewall rules, do I have to open specific ports additionally somewhere? just to ensure that I am doing it correctly...
Service All basically means every port. Effectively allowing Service All means the firewall starts allowing everything and no longer protects as a firewall should.
Useful for debugging, horrible for security. Its like getting a really expensive deadbolt for your front door and then leaving your door wide open. :)
Thank you, then I'd been right. For testing purposes I moved all WiFi Clients into a separate VLAN with its own Firewall rule on that VLAN. Therefore I allowed all services so unfortunately this does not seem to be my solution either.
Michael, do you think this could be related to my VLAN Setup? According to my logs I also had some default Drops (rule ID 0) with "Could not assocate packet to any connection"
Thank you for everthing so far.
Additionaly: I am testing with our Aruba Access Points at the moment, using a SSID that does not support Roaming accross other APs. Maybe something happens to the session during roaming and Sophos stumbles upon that. Just to exclude the Aruba I'll give that a try. Keep you updated.
My recommendation from Sept 1 still stands.
1) You have done a bunch of work to diagnose. It is very hard for people in the forums to diagnose. It is easier for Sophos Support to since you can give them direct access to your system. Raise a ticket and pester them.
2) Use firewall rules to bypass the proxy just for these connections. Create FQDN Host objects for all of the whatapp domains and then put in a high level rule the has them as the destination, service any. Traffic that is not for WhatsApp still uses your regular firewall rule with web proxy.
I finally, after months, solved the issue:
Go to Web Protection -> Generall Settings -> HTTPS decryption and scanning and then UNCHECK "Block unrecognized SSL protocols"
Somewhere I read about Whatsapp using invalid SSL. Thats the point here. Wished my firewall had more clear logfiles. Impossible to work with the logfile to solve this. Thats the sad point. The lucky point is, it is working :) :) :) Finally! ;)
Works with pharming protection on, btw. And I dont need any exceptions. You can leave "Block invalid certificates" checked.