Problem when hosting with incoming https traffic

rlyness · ‎20-04-2026

I've hosted a football guessing website for a number of years on a dedicated server in my house (just for my friends for some fun). I've had no problems until a few weeks ago when I renewed my contract with EE broadband.

Now, I am getting weird intermitting problems with some of the guys trying to connect to my website. When they're on their home internet (different providers), they get a "Site cannot be reached ERR_TIMED_OUT" error - but when they drop off wifi and go on mobile internet, it magically just works. To add, sometimes when I restart the router it suddenly works for a while then stops working again.

I've got a dynamic DNS set on my router so I know it's not a problem with routing. It just, for some reason is not responding while on their home WIFI. Also, it's not all users - some can access fine both on home WIFI and mobile. It only seems to be https traffic affected, coming through on http fine all of the time. My certs are valid, my port forwarding hasn't changed for years so I know all that is setup fine.

I've done a full factory reset of the router, I've temporarily hosted the site on a different machine on the network. I'm all out of ideas. Anyone had any similar experience?

bobpullen · ‎18-06-2026

@rlyness - thanks for updating. All very peculiar though! I know there were some PSU issues with some of the early Smart Hub 6 Plus devices but I'd be amazed if that were to selectively affect port forwarding.

All's well that ends well though, I guess 🙂

rlyness · ‎18-06-2026

@bobpullen yes, all very strange. Technician ran a service test and said there were loads of service alerts flagged on the router so definitely something was going on there.

Also wondering if new router will solve this problem that I've also been having: https://community.ee.co.uk/t5/Broadband-Landline/Microsoft-Teams/m-p/1582599#M129980

I'll be working from home tomorrow so looking forward to testing that out.

Again, thank you (and all other replies) for all your suggestions here. Very much appreciated

JimM11 · ‎18-06-2026

@rlyness Can you do a quick check on your EE Smart Hub 6+, take a look at the current FW, as new it may have just popped onto the latest when first connected?

rlyness · ‎18-06-2026

Hi @JimM11 - I'm in the office today but will have a look when I get home. When I plugged it in, it did take a few minutes to come to live and seemed to restart once or twice so assume it was doing some kind of update. Will reply later today.

JimM11 · ‎18-06-2026

@rlyness Thanks for the reply, and yes correct normal OOB FW updates get applied, tries hard to do all the account matching, keeping the same SSID/Password etc so as not to have you doing the hair pullout and lot's off sweary words. How the versions have went below!

FW: r3.8.30-R-1627300-PROD-83002 and the gui is at App version 3.1.18 05/12/2025
FW: r4.26.1-R-1860948-PROD-83002 and the gui is at App version 3.14.5 25/05/2026 OP Posted Stability Problem
FW: r4.26.3-R-1923144-PROD-83002 and the gui is at App version 3.14.5 12/06/2026 OP Posted

bobpullen · ‎20-06-2026

@rlyness wrote:
Also wondering if new router will solve this problem that I've also been having: https://community.ee.co.uk/t5/Broadband-Landline/Microsoft-Teams/m-p/1582599#M129980

See here. In particular, this bit: -

Fix - Microsoft Teams/UDP issues in some situations

So, hopefully fixed once your router is upgraded to a 4.26.x firmware 🤞

rlyness · ‎20-06-2026

Hi @bobpullen

Thanks again for the helpful reply. I did a teams call yesterday all did seem to work well actually.

However, after ~70 hours of my website working fine, it started to fail again. I downloaded 5600 IIS logs and fed them into Claude. It suggested that an "MTU black hole"

It suggested the fix would be to lower the MTU setting on the router - but I couldn't find this setting on my SmartHub 6. It said I could set this setting on the network adapter on the server in which my site is hosted using PowerShell.

And.. to my amazement - it just worked.

I'm fairly sure something changed in the FW somewhere as this has worked for many years and suddenly stopped.

bobpullen · ‎20-06-2026

@rlyness - I almost let out an audible groan before I got to the bit about the MTU fix 😉

Remind me, did you change access technology when you recontracted? e.g. from part fibre/FTTC > Full Fibre?

The EE hubs don't allow tweaking of the MTU, but from what I can tell it's set to 1492, which is expected for a PPPoE connection on the Openreach network.

Below shows things fragmenting above 1464 bytes. 1464 + 28 (for IP/ICMP overhead) = 1492: -

PS C:\Users\Bob> ping www.google.com -f -l 1464

Pinging www.google.com [142.251.152.119] with 1464 bytes of data:
Reply from 142.251.152.119: bytes=1464 time=10ms TTL=117
Reply from 142.251.152.119: bytes=1464 time=10ms TTL=117
Reply from 142.251.152.119: bytes=1464 time=10ms TTL=117
Reply from 142.251.152.119: bytes=1464 time=9ms TTL=117

Ping statistics for 142.251.152.119:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 9ms, Maximum = 10ms, Average = 9ms

PS C:\Users\Bob> ping www.google.com -f -l 1465

Pinging www.google.com [142.251.157.119] with 1465 bytes of data:
Reply from 192.168.1.254: Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.

Ping statistics for 142.251.157.119:
    Packets: Sent = 4, Received = 1, Lost = 3 (75% loss),

JimM11 · ‎20-06-2026

@bobpullen MTU's the bain off ones life, how windows ping see's it also.....

C:\Windows\System32>ping www.google.com -f -l 1465

Pinging www.google.com [142.251.150.119] with 1465 bytes of data:
Reply from 192.168.50.1: Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.

Ping statistics for 142.251.150.119:
Packets: Sent = 4, Received = 1, Lost = 3 (75% loss),

C:\Windows\System32>ping www.google.com -f -l 1464

Pinging www.google.com [142.251.150.119] with 1464 bytes of data:
Reply from 142.251.150.119: bytes=1464 time=21ms TTL=119
Reply from 142.251.150.119: bytes=1464 time=20ms TTL=119
Reply from 142.251.150.119: bytes=1464 time=20ms TTL=119
Reply from 142.251.150.119: bytes=1464 time=18ms TTL=119

Ping statistics for 142.251.150.119:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 18ms, Maximum = 21ms, Average = 19ms

C:\Windows\System32>

My Asus currently PPPoE on vodafone set itself up as MTU = 1492 when adjusted and hung direct on the ONT....

Was on sky connection before so DHCP_PD6 set routing operation none off there MER 61 at all....

rlyness · ‎20-06-2026

@bobpullen I was full fibre with BT before EE poached me to change. As far as I knew it was a like for like swap. Would need to look through mails to see if there was any mention of the technology

My ping to google is different to yours. Although, I accidently ran the PowerShell on my personal PC (and not the server) yesterday so now my ping google looks like this:

Fine at 1424+28=1452
Frag at 1425=28=1453

-------------

Z:\>ping google.com -f -l 1424

Pinging google.com [142.251.29.100] with 1424 bytes of data:
Reply from 142.251.29.100: bytes=1424 time=16ms TTL=113
Reply from 142.251.29.100: bytes=1424 time=15ms TTL=113

Ping statistics for 142.251.29.100:
Packets: Sent = 2, Received = 2, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 15ms, Maximum = 16ms, Average = 15ms
Control-C
^C
Z:\>ping google.com -f -l 1425

Pinging google.com [142.251.29.100] with 1425 bytes of data:
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.

Ping statistics for 142.251.29.100:
Packets: Sent = 2, Received = 0, Lost = 2 (100% loss),
Control-C
^C
Z:\>

----

EDIT: Setting my personal PC back to 1500 (windows default?) I get the same replies as you. From my server, now that I set the MTU to 1452, i get a frag at 1425