Internet Outage in November. Poor Connection ever since.

[#1]

Thanks for getting back to me @Paula7 and apologies for the delayed response.

I've left the Qube engineer notes on to the fault ticket. I'm afraid that Qube engineers can't investigate a fault externally, for that we'd need to arrange an Openreach engineer and you can arrange this by replying to the ticket logged Here. because I can't book an engineer directly from a request on our Community Forums.

However, if you can let me know here once you've responded to the ticket, I'll make sure we book it as soon as we can.

Regarding the fault closure, this was closed in error and I'm really sorry for the inconvenience. 

[#2]

So I've run RouterStats on and off over the past few days and unfortunately nothing really stands out for me from the monitoring...

When the websites are connecting (and when they are not) - there is no variation in the stats.

I've attached the Helpdesk stats from today, and the charts from RouterStats from Friday (when the internet was hardly working at all)

A couple of things from the router event log - incase they mean anything - this CWMP timeout happens regularly.

13:50:33, 14 Feb.	(357803.870000) CWMP: session closed due to error: Timeout
13:50:03, 14 Feb.	(357773.440000) CWMP: Server URL: https://dbtpnhdmw.bt.mo; Connecting as user: Default username
13:50:03, 14 Feb.	(357773.430000) CWMP: Session start now. Event code(s): '0 BOOTSTRAP,4 VALUE CHANGE,2 PERIODIC'

and again 15 mins later...

14:06:04, 14 Feb.	(358734.730000) CWMP: session closed due to error: Timeout
14:05:34, 14 Feb.	(358704.190000) CWMP: Server URL: https://dbtpnhdmw.bt.mo; Connecting as user: Default username
14:05:34, 14 Feb.	(358704.180000) CWMP: Session start now. Event code(s): '0 BOOTSTRAP,4 VALUE CHANGE,2 PERIODIC'

I also get quite a lot of these messages..

13:36:44, 11 Feb.	IN: BLOCK [12] Spoofing protection (IGMP 31.185.211.178-​>224.0.0.22 on ppp0)
13:36:26, 11 Feb.	BLOCKED 4 more packets (because of First packet is Invalid)
13:36:25, 11 Feb.	OUT: BLOCK [65] First packet is Invalid (Invalid tcp flags for current tcp state: TCP [192.168.1.64]:60483-​>[204.79.197.200]:443 on ppp0)

Many thanks in advance to anyone who has any further suggestions as to what the issue could be.

@Gandalf  - I've responded to the ticket with availability for the Openreach Engineer.  Thanks for the apology.  Good to see that the Qube engineer did confirm the problem.

[#3]

@Paula7 I haven't read back on this topic before posting this, but the Help Desk report makes satisfactory reading - you are still on an old software release, although I understood all Hub1s had now been updated to 289.1.3 release. What is your MGALS figure ?

The RS graphs look OK to me - in fact I wish mine were as steady when I was on ADSL with PN. Those Event log issues are nothing to be really concerned about, although I am left wondering if those CWMP failures are the result of the system trying to update the firmware and failing, for some reason?

[#4]

@jab1 Thanks for taking a look.  My MGALS figure is at 1.7Mbps.

All the speed tests show I am getting much better speed than that... its not a speed issue, everything points to the line being fine.

Clutching at straws now ... but could there be a setting somewhere in plusnets system, that is not recognising my connection somehow (perhaps username or ip address).  Clicking on websites seem to timeout, as the CWMP failure in the event log so maybe the issues are connected?

Random thought 1 - could someone else be sharing my access somehow - maybe someone else has been set up with the same username (would have thought impossible but...).  A couple of times I have got a message saying another device with the same name was using the network (I can't remember what I was doing at the time) .  I ignored it at the time as there are no devices with even similar names on my home network.

Random thought 2 - Plusnet advisers keep telling me that they have phoned me back on my landline but I have always been sat in the same room as the phone and it never rings (and doesn't show any missed calls).  I have similarly ignored this as potentially poor service, but the last time, the adviser also called me on my mobile, and was adamant he had called the landline first...  I don't have any other issues with my phone - other people are able to call.

So, just a thought if they are phoning using their system  (and my account is somehow set up wrong?) then maybe they can't get through to me - in the same way my internet doesn't appear to connect to plusnet?

As I said clutching at straws with random non-technical thoughts... but maybe could happen?

[#5]

Random thought 1 - I very seriously doubt that - BB provision is based on phone numbers, i.e. your unique identifier on the BT system.

Random thought 2 -  Linked with (1) above - check your account pages, the phone number used for supplying your BB is on there somewhere - can't tell you exactly where now as I'm not with PN any more, so don't have full account detail access.

I would possibly try ringing in and asking what landline number they have for you?

[#6]

We are looking for transient events.  Those stats do not have enough granularity.  You need a 10 second sampling interval and a plot which shows 2 hours of stats - that is 1,200 data points.

[#7]

@Townman Hadn't noticed the scale - sorry.

[#8]

So today, Openreach came and checked the line.  They found no problems.  All checks were passed.

Showed the engineer the issues on my wired pc and (unprompted from me), he thought it was an issue connecting with Plusnet...at their end.

@Townman I've re-run the router stats at 10sec intervals and 1200 data points.  Hope I got it right this time.

Charts attached. The drop at 15.30ish on the noise margin was a phone call received on the landline.

Same connection problems throughout the timeframe of the 2 hour charts.

[#9]

@Paula7 Those RS graphs look OK to me, BUT it may be an issue with the phone side of the connection if a call causes a blip like that.

@Townman Any ideas?

[#10 · Accepted solution]

---

@Paula7 wrote:
The drop at 15.30ish on the noise margin was a phone call received on the landline.

---

Assuming all’s checked internally (tried the test socket, no extension cabling between the router and telephone socket, replaced microfilter etc) this fault condition seems to suggest a faulty line card at the exchange, which is something an Openreach engineer will need to check and carry out a “lift & shift” to a different card. 

[#11]

@jab1  Hi John,

Thanks for your views.  Do you have any alternative ideas as to the problem?

At the start of this issue, I would have agreed these random thoughts were impossible but I feel that I have exhausted all alternatives so now looking at the highly unlikely.

I am aware that BT uses phone numbers as the unique identifier, but each internet provider requires a username/password to access the line.  It is this verification step with Plusnet that I am thinking could be the problem in some way. The sharing suggestion is due to the intermittent nature of the issue. If the username/pwd was wrong it would obviously not work at all.

I already asked about the phone number they hold for me a while back.  The adviser laughed and said they provide the line. I was told they call through their system and don't need to refer to the contact details (BTW the contact details are correct on the account).

[#12]

@Anoush 

Yes, the engineer worked from the test socket and used his own microfilter.

If phone calls are the issue, why is the connection so bad all day every day?

The phone call today made no difference to the issues connecting to websites.  It has been bad all day.

[#13]

Paula - sorry been out all day, so I'm a little frazzled at the moment.😉 I'll have a re-read of the whole topic first thing tomorrow when my head is a little clearer, but too be honest nothing springs to mind, which is one of the reasons I tagged Townman above.

I still doubt it is anything to do with your Plusnet account verification, otherwise I don't think you would connect at all, and PN usernames/passwords are tied to your account and cannot be used by other customers.

If your contact details are correct on the account something tells me the agent was being a little 'flippant', to say the least.

[#14]

Thanks for getting back to me @Paula7 

I’m not suggesting the issue’s just present when you make or receive a phone call, but if the noise margin spikes when that happens, then it’s very possible that the exchange equipment has developed a fault, which can affect your broadband, even if a call isn’t made. 

Basically, faulty equipment can manifest itself in various ways, and seeing what a phone call does to your broadband may ultimately have diagnosed the issue. 

I’m not in the office now until Friday but one of my colleagues should hopefully be able to pick this back up for you in the meantime. 

[#15]

The noise plot is the only one of relevance. The plot you’ve shared somewhat points to something not right on the line / exchange equipment. The more granular plot identifies problematic transient events which can cripple internet stability, yet not show up as a line fault when tested.

@Gandalf ’s approach is spot on, given all of the evidence available. Keep the monitoring running and look at the plots at the times you perceive issues. Set up router stats to load when the computer starts up … and if you can, leave running 24x7.