Potential STP (spanning tree) problem with Hub 2 when playing with some wirelessly connected devices

[#1]

WAN access to resources outside the LAN affected. Connecting to the modem itself on the LAN affected.

I've tried my best to to do troubleshooting in a LAN with limited resources. But the symptoms certainly look like a STP problem.

It is possibly related to Wifi connected devices. Possibly with Samsung Galaxy smartphones.

Problem only seen when a Galaxy connected to the LAN.

At the first hint of a new occurence of this problem I switched the hub to wireless mode 'mode 2'. The side effect was that ICMP pings to a Galaxy were suddenly in the 10 or less ms bracket. Whereas they were not before (anything up to 200ms or thereabouts).

Just my thoughts/experiences.

[#2]

Aside. Galaxys don't really do MAC spoofing. Troubleshooting would be easier with MAC filtering.

[#3]

Troubleshooting what exactly? What specific problems are you are experiencing, how does the issue manifest itself, and what exactly are you doing to trigger it?

At the risk of sounding blunt, these posts are akin to saying - "There is a problem with 'X' which would be easier to troubleshoot if I had 'Y'" - It doesn't provide much to work with if your intention is for us to look at fixing something

[#4]

Ignore me. I've now reset to factory defaults. Once I get to have a look at other connected devices I'll have my suspicions confirmed on where the problem may lie. It is intermittent.

Hub configured as DHCP server (it makes it easier for the kids when they visit as all they need is the wireless key)

The main (most worrying) symptom has been (from a cabled destop)

No response from web browsing (even http://192.168.1.254)

Pings timing out (but not always)

If we can leave it there for the moment I'm having a delve around. It is irritating me. I refuse to be beaten

Just one thing if anyone can positively confirm.

If the hub is put into bridge mode does it act purely as a bridge, i.e. are all wireless functions disabled and all wireless transmissions disappear (i.e. zero possibility of interference with other wireless)?

(Just so I know when I go shopping for an ethernet connected router)

[#5]

Spanning tree problems :

Not seen any of those since working on bonded kilostream lines in the mid 80's when packets would go missing in the network...

[#6]

---

@greygit wrote:

Just one thing if anyone can positively confirm.

If the hub is put into bridge mode does it act purely as a bridge, i.e. are all wireless functions disabled and all wireless transmissions disappear (i.e. zero possibility of interference with other wireless)?

---

Yes, it will act purely as a modem.

[#7]

This answer your question?

[#8]

Update:

I was hoping that a reset to factory defaults would uncover/undo some terrible blunder I had made. But, no. It appears that my original thoughts were correct. I'm currently having to assume that the TG582n (with a failing microswitch) had some underlying QoS which had previously hidden this. Or it is something that has happened when the Hub2 went in (highly unlikely IMO).

With a little more thought from me (rewind my brain over a decade and desperately try to remember basic IP level troubleshooting stuff without using anything more sophisticated than ping and telnet) I have now come full circle and have arrived back at what I thought originally. There's a local device that is occasionally totally saturating the upstream bandwidth. To the obvious detriment of other local devices.

My original thoughts were based on upload/download figures as recorded by the Hub2 for individual devices. Hence my original 'interest' in the ability to reset them without a reboot to validate what I thought was happening. But, hey. Get the pencils out and re-sharpen them.

When pings to the PN dns server are nearly 2 seconds then it isn't really surprising that DNS resolutions time out and stuff stops working.

Any chance of QoS settings on the hub for local devices? Pretty please Until the (potentially rogue) app on a phone which isn't mine has been identified and has been effectively neutered. It only happens periodically. Let's just say knocking the device off the web permanently is not conducive to marital relationships. T'other 'alf then just uses 4G and uses up allocated quota and then moans at me when there are further financial costs.

I have residual questions about the DHCP implementation on the hub2 (parked on the back burner for now).

Could I upgrade the line to accomate the use? As I've previously alluded, the bit (for some value of measurement) of slightly damp copper (underground, overground, swinging in the wind and perched on by birds) which connects me doesn't provide real benefits if I upgrade to BT's VDSL. The BT checker is somewhat circumspect in its estimations of what I would see. Estimated upstream bandwidth is not much greater than what I get.

[#9]

Latest update. The 'offending' smartphone doesn't have to be in active use. Just connected to WiFi. Stuff going on in the background, happily using all available upstream bandwidth. Since that device only copes with 2.4GHz that part of the wireless network is now disabled.

I suppose any mobile network operator would be happy when an app uses up their subscribers monthly allowance and they can get new income.

Yes, that's a little unkind, but it appears that there's a lot of less than optimal (overly polite phrasing IMO) mobie apps out there. I've been lurking on one or two mobile phone app developer places and the level of cluefulness is woeful. IMO.

[#10]

There's a multitude of stuff which runs in the background, both native to the phone's OS and added applications.

Which phone and what apps?  Others might have helpful experience.

[#11]

"There's a multitude of stuff which runs in the background, both native to the phone's OS and added applications."

Yes. I know at least this much.:) Given a choice I'd be doing a proper reset to factory default, but that results in the possible loss of data/config.

[#12]

That's (also) what the Hub 2 config/display looks like when the WiFi is switched off and it isn't configured as a modem.

Did that (fully) anwer my question? I'll leave that for you to decide.

[#13]

Assuming I have followed this thread correctly it seems that you have ADSL rather than VDSL (“fibre” [to the cabinet] broadband).

When I had ADSL I found that uplink saturation was the thin that killed broadband performance (I had downlink speed available) and that was what sold me on VDSL.

I tried applying a QoS policy (not awfully hard) but the thing that worked best was making the WiFi as slow as possible (thus discouraging uploads) and turning off Microsoft One Drive.

In my case I had decent VDSL speeds so that fixed it. In your case it sounds like VDSL won’t help so much but any upload improvements may be worthwhile.

I also assume you have tried everything with internal wiring improvements.

[#14]

BT's assessment of the line attached. I'd anticipate that my line is assessed by BT as 'impacted'.

FTTPoD is apparently available, but FTTP is not available. That's interesting. The nearest fibre install is measured in miles. I suspect there's a wriggle clause or three that BT know about when it comes to FTTPoD,

Your comments re: saturation of upstream/uplink is correct. If traffic can't get out (gets dropped), then replies will not be coming back in. QoS on the local LAN has to have a modem/router that understands it. QoS settings on a customer's local connection/traffic will/should be ignored once the traffic exits into the upstream. Unless there is a financial arrangement between the customer and their upstream.

[#15]

@markhawkin The Broadband checker result provided after your post is slightly out of date - 2020, and I think the OP has said elsewhere that they are now on FTTC, although that is still providing an extremely low estimated max speed.