General Discussion
  >> Fibre Broadband


Register (or login) on our website and you will not see this ad.


Pages in this thread: 1 | 2 | >> (show all)   Print Thread
Standard User Andrue
(knowledge is power) Sun 21-Apr-13 10:48:37
Print Post

The 'guess that fault' thread.


[link to this post]
 
Connection running normally:

http://www.thinkbroadband.com/ping/share/33c68fedd7d...

Fault graph:

http://www.thinkbroadband.com/ping/share/378fe4a258e...

Comments?

---
Andrue Cope
Brackley, UK

Just because he could. RIP.

Edited by Andrue (Sun 21-Apr-13 10:49:32)

Standard User Chrysalis
(legend) Sun 21-Apr-13 10:54:23
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
well for a start I think you have jumped the gun to say it cant be the copper.

BT Infinity 2 Since Dec 2012
Standard User R0NSKI
(experienced) Sun 21-Apr-13 11:02:26
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
To me it just looks like there is activity on the connection, causing the spikes in latency. My connection gets a lot of yellow spikes when my cloud backup is backing up.

But IIRC your speed tests reduce greatly when the problem occurs.

Have you tried disconnecting your home network/turning off wireless when it occurs to see what affect it has on the ping graphs?

Edited by R0NSKI (Sun 21-Apr-13 11:03:58)


Register (or login) on our website and you will not see this ad.

Standard User Andrue
(knowledge is power) Sun 21-Apr-13 11:59:38
Print Post

Re: The 'guess that fault' thread.


[re: R0NSKI] [link to this post]
 
In reply to a post by R0NSKI:
To me it just looks like there is activity on the connection, causing the spikes in latency. My connection gets a lot of yellow spikes when my cloud backup is backing up.

But IIRC your speed tests reduce greatly when the problem occurs.
Yeah, as usual roughly 50% drop in throughput. ISP status page shows that yesterday was actually an unusually low day for usage - 210MB downloaded.

---
Andrue Cope
Brackley, UK

Just because he could. RIP.
Standard User Andrue
(knowledge is power) Sun 21-Apr-13 12:01:19
Print Post

Re: The 'guess that fault' thread.


[re: Chrysalis] [link to this post]
 
In reply to a post by Chrysalis:
well for a start I think you have jumped the gun to say it cant be the copper.
Lol. The modem isn't showing anything.

xdslcmd info --stats
xdslcmd: ADSL driver and PHY status
Status: Showtime
Retrain Reason: 2
Max: Upstream rate = 26634 Kbps, Downstream rate = 76644 Kbps
Path: 0, Upstream rate = 20000 Kbps, Downstream rate = 65847 Kbps

Link Power State: L0
Mode: VDSL2 Annex B
VDSL2 Profile: Profile 17a
TPS-TC: PTM Mode
Trellis: U:ON /D:ON
Line Status: No Defect
Training Status: Showtime
Down Up
SNR (dB): 5.9 15.0
Attn(dB): 0.0 0.0
Pwr(dBm): 12.5 6.4
VDSL2 framing
Path 0
B: 63 237
M: 1 1
T: 64 45
R: 16 16
S: 0.0309 0.3782
L: 20698 5373
D: 1043 1
I: 80 127
N: 80 254
Counters
Path 0
OHF: 598902396 750065
OHFErr: 206 522
RS: 3320109961 1801097
RSCorr: 1268545 3759
RSUnCorr: 10342 0

Path 0
HEC: 3073 0
OCD: 0 0
LCD: 0 0
Total Cells: 1275389047 0
Data Cells: 451167261 0
Drop Cells: 0
Bit Errors: 0 0

ES: 55278 2350
SES: 48 0
UAS: 51 51
AS: 892383

Path 0
INP: 3.00 0.00
PER: 1.48 4.25
delay: 8.00 0.00
OR: 129.36 60.17

Bitswap: 116597 5829

Total time = 1 days 43 min 13 sec
FEC: 2948848 24782
CRC: 3760 2539
ES: 55278 2350
SES: 48 0
UAS: 51 51
LOS: 9 0
LOF: 10 0
Latest 15 minutes time = 13 min 13 sec
FEC: 1790 4
CRC: 0 2
ES: 0 1
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Previous 15 minutes time = 15 min 0 sec
FEC: 2362 3
CRC: 0 0
ES: 0 0
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Latest 1 day time = 43 min 13 sec
FEC: 4758 11
CRC: 0 2
ES: 0 1
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Previous 1 day time = 24 hours 0 sec
FEC: 110364 365
CRC: 13 65
ES: 6 60
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
Since Link time = 10 days 7 hours 53 min 3 sec
FEC: 1268545 3759
CRC: 206 522
ES: 38 488
SES: 0 0
UAS: 0 0
LOS: 0 0
LOF: 0 0
#

---
Andrue Cope
Brackley, UK

Just because he could. RIP.
Standard User deleted
(deleted) Sun 21-Apr-13 15:51:24
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
On the fault graph, the symptoms you see are some very minor yellow peaks - showing that the maximum latency gets rather higher than normal, but not in any particularly continuous way, and with no special change during the evening peak hour

For comparison, here's mine. Note the evening peak - which is down to traffic at the ISP or on the WBC backhaul, not my link at all.

Note, especially, that your blue areas *don't* change in the same way as the yellow. That means the delays are happening to individual packets, on an occasional basis, keeping the average delays low.

[Detailed maths: My estimation is that your yellow mini peaks represent about 15ms delay, and that 1 horizontal pixel represents 100 seconds. That means at least 1 ping in 100 is getting delayed by 15ms, and the average additional delay is 0.15ms. When it gets to 7 pings in 100 delayed by 15ms, the average increases to 1ms - and the blue line moves 1 pixel higher.]

So yellow peaks alone mean 1-3 packets in 100 are being delayed by 15ms. Yellow peaks with a single extra pixel in blue means around 7 packets in 100 are being delayed by 15ms.

Conclusion: Every ping packet is getting to the modem, and 95%+ are getting back in the minimum latency. Even when there are delays, it only affects 10% or so of packets, by up to 15ms.

Like R0nski, I think this suggests that something else is transferring, that causes the ping packets to be queued slightly. But the queue could be happening *anywhere* between TBB and your router (or whatever you have responding to the ping).

The one place that wouldn't queue your ping packets is the modem (your modem or the cabinet's modem), at least due to errors: Errors in the modem-modem stream cause packets to be dropped, never queued. The modem probably could queue the ping due to QoS, but that would be triggered by some other download/upload happening.

Note, especially too, that there are no red areas - there is zero packet loss. Because ping packets are never retried (by design), and you can see zero loss, you know for certain that they aren't being lost on the modem-modem link either.

No red areas also means no resync's, for any reason at all. DLM hasn't intervened to turn anything on or off, and there has been no independent loss of signal.

On that evidence, there is *nothing* wrong with the copper.

There could be something wrong with your modem, that make it *look* to be working perfectly. For example, that it is responding perfectly to pings but choosing to delay or drop your downloads, but that would be a very individual fault. Seems unlikely to me.

HOWEVER

on your "working normally" graph, you get a blip of reduced latency.

There is no packet loss, so there have been no resyncs to cause this. Either we blame this on a brief re-routing of packets within BT's 21CN WBC network, or we suspect something with your modem.

It is hard to believe that your modem has a fault that just happens to manifest itself by appearing to work perfectly on the TBB Ping graph.
Standard User Andrue
(knowledge is power) Sun 21-Apr-13 16:19:20
Print Post

Re: The 'guess that fault' thread.


[re: deleted] [link to this post]
 
In reply to a post by WWWombat:
It is hard to believe that your modem has a fault that just happens to manifest itself by appearing to work perfectly on the TBB Ping graph.
Nor that manifests every 60 days starting at 1500utc:

April (4pm BST)
http://www.thinkbroadband.com/ping/share/1ae25650a99...

February (3pm GMT)
http://www.thinkbroadband.com/ping/share/1b414a57412...

Just have to wait for proper Support staff on Monday to where they go next with this. If nothing much happens I may switch to AAISP. Such a repeatable must be fixable if only we can find the right BT bloke to talk to :-/

---
Andrue Cope
Brackley, UK

Just because he could. RIP.
Standard User deleted
(deleted) Sun 21-Apr-13 16:59:37
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
Something behind the scenes, almost certainly!
Standard User Andrue
(knowledge is power) Mon 22-Apr-13 20:09:28
Print Post

Re: The 'guess that fault' thread.


[re: deleted] [link to this post]
 
In reply to a post by WWWombat:
Something behind the scenes, almost certainly!
Cleared now. BT apparently claim to have cleared a 'profile mismatch' over the weekend. I didn't know until Support got in this morning and there was no sign of a change. However this evening in an attempt to further diagnose I disconnected my router and therefore got a new session. Hard to say if it's just cleared because it's due or because of what BT did.

http://www.thinkbroadband.com/ping/share/8c5c9fd6841...

Modem has been up all through this :-/

---
Andrue Cope
Brackley, UK

Just because he could. RIP.

Edited by Andrue (Mon 22-Apr-13 20:11:29)

Standard User deleted
(deleted) Tue 23-Apr-13 00:15:04
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
Graph fails to load here.
Standard User Andrue
(knowledge is power) Tue 23-Apr-13 08:03:32
Print Post

Re: The 'guess that fault' thread.


[re: deleted] [link to this post]
 
In reply to a post by WWWombat:
Graph fails to load here.
Strange, seems to be fine from work. Anyway there's nothing much to see, just a flat graph with a few spikes courtesy of my Samknows box.

---
Andrue Cope
Brackley, UK

Just because he could. RIP.
Standard User deleted
(deleted) Tue 23-Apr-13 12:43:51
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
OK - back to the original state then.

I wonder how a profile mismatch can cause this kind of symptom (ie low level packet queues and delays) with the sort of regularity you've seen.

I can't imagine that the profile has anything to do with L2 routing within the BT network, but perhaps they pointed you via a strange node, or backup leg.
Standard User Andrue
(knowledge is power) Tue 23-Apr-13 13:48:10
Print Post

Re: The 'guess that fault' thread.


[re: deleted] [link to this post]
 
In reply to a post by WWWombat:
I wonder how a profile mismatch can cause this kind of symptom (ie low level packet queues and delays) with the sort of regularity you've seen.

I can't imagine that the profile has anything to do with L2 routing within the BT network, but perhaps they pointed you via a strange node, or backup leg.
It sounds odd to me as well. I was a bit surprised that it was fixed after yesterday's login but looking back it seems that the August session also cleared on the 22nd. So whether it had anything to do with that or not I don't know.

Meanwhile it appears my laptop has decided that it doesn't want to support PPPoE. No idea what was going on there. I just tried to use the existing network connectoid and all it did was jam the hard drive light on and slow the machine to the point where it locked up. I don't think it ever actually established a connection.

---
Andrue Cope
Brackley, UK

Just because he could. RIP.
Standard User Andrue
(knowledge is power) Thu 11-Jul-13 07:45:42
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
I forgot I hadn't updated this thread. Well - the news was that June came and went and my connection has been fine. There has been one change though. I'm now using a different router. The only problem with that as an explanation is that I have previously eliminated the router as a cause by showing that the fault exists when my laptop is handling PPPoE.

But on the face of it it does look like my old wnr1000 might have been the cause. But what could it have been doing that created a fault that lasted four or five days and survived the router being powered off (for over 20 minutes) and a different device (latop) handling login? The fault even persisted on a temporary login that my ISP gave me.

Weird.

---
Andrue Cope
Brackley, UK
Standard User Andrue
(knowledge is power) Sun 27-Oct-13 18:38:57
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
Well here's a thing - I've been contacted by someone who's seeing the same thing on PN:

http://community.plus.net/forum/index.php/topic,1193...

Interestingly in discussions with him here it seems he has the same router that I used to use. It's definitely looking like the router is a key component. It makes sense on one level - the fact that the modem never blinked - but there's still the fact that it takes several days before the fault goes away.

---
Andrue Cope
Brackley, UK

Edited by Andrue (Sun 27-Oct-13 18:53:33)

Standard User deleted
(deleted) Wed 30-Oct-13 15:00:05
Print Post

Re: The 'guess that fault' thread.


[re: Andrue] [link to this post]
 
Hi All,
Well, a Bt man came this morning, as arranged, and tested out the line. He tested the sync speed and got about 73Mb. But could see that when we did a BTW speed test, the data transfer rate was only around 30 to 35Mb. Exactly as I had found, and confirming Plusnets own tests.
He wasn't aware of any scheduled events at the excahnge that happened every 2 months and was basically stumped.

We decided to try a new BT Homehub4 and an ECI based modem that he had on the van, in place of my Netgear WNR1000 router and Huwai modem, this gave a speedtest result of 71Mb! So for completeness, we then put my Netgear back on with the ECI modem, and again got nearly 70Mb. Finally with my router and original modem the speed was 63Mb. i.e. everything working like it should!

Once before in one of the previous slow downs, (when it is only a day away from the fault fixing itself on the first of the month)
I had rebooted modem and router and resolved the problem. But if I try and do that a day or so after the fault first appears, then it won't fix it. So I'm not entirely sure if putting a new modem on was the cure or not.

What I intend to do, is source a different router and leave that on for the next possible occurence, which should be December 18/19th, then if the problem doesn't happen, put the Netgear back on again and see if the fault comes back in February.
This should prove one way or the other if the Netgear is somehow responsible.

As the Netgear WNR1000 was issued by Plusnet as part of the original FTTC trials, i wonder if any of the other trialists had speed problems and just changed out the router at the first occurence, and so had not seen it would happen every 2 months??
Standard User RobertoS
(sensei) Wed 30-Oct-13 15:22:13
Print Post

Re: The 'guess that fault' thread.


[re: deleted] [link to this post]
 
Ah, the Kevin that Andrue linked to smile. That was a bit confusing at first tongue. I thought Andrue had a new nick for some reason.

Certainly a vey weird fault.

My broadband basic info/help site - www.robertos.me.uk | Domains,website and mail hosting - Tsohost.
Connection - Plusnet UnLim Fibre (FTTC). Sync ~ 55.8/14.5Mbps @ 600m. - BQM

"Where talent is a dwarf, self-esteem is a giant." - Jean-Antoine Petit-Senn.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Allergy information: This post was manufactured in an environment where nuts are present. It may include traces of understatement, litotes and humour.
Standard User deleted
(deleted) Sat 21-Dec-13 15:49:54
Print Post

Re: The 'guess that fault' thread.


[re: RobertoS] [link to this post]
 
Well, the 18th/19th has come and gone, and my speed is still solid at 59Mb according to SamKnows.
To see if the Netgear router is definitely the culprit, I'm going to put it back on in the New Year and wait for Feb 18th.
If it does cause a slow, I'll then know for certain *what*, but I still won't have a clue *why*! smile
Standard User deleted
(deleted) Sat 21-Dec-13 18:09:25
Print Post

Re: The 'guess that fault' thread.


[re: deleted] [link to this post]
 
In reply to a post by WWWombat:
Graph fails to load here.


Same here.
Pages in this thread: 1 | 2 | >> (show all)   Print Thread

Jump to