Kitz Forum
Broadband Related => ADSL Issues => Topic started by: Weaver on December 06, 2022, 07:55:26 PM
-
Some weeks back I got some rather upsetting emails from A & A about the problem I was having with my then HCD-inflicted line 2. Line 2 has since recovered on its own and is now perfect. Instead, line 4 has now caught extremely bad HCD and is almost permanently down now, but when it’s up it’s unusable, with so many errors and such a slow downstream rate that it needs to be disabled. In respect of line 2 AA said something that was a bit confusing but it sounded as if possibly AA had run into big trouble with either BTW and/or Openreach and one or the other was I assume refusing to fix the fault but I couldn’t make it out. Anyway, AA said they had reached the end of the road and couldn’t fix line 2. Today’s email, surprisingly, and disappointingly seems to extend this to line 4 and they’re refusing to fix line 4’s HCD either. Perhaps I’m supposed to be grateful that they haven’t just killed the line 4, or worse killed all of my DSL lines.
I realise that AA isn’t even remotely in control of this and are always reliant on this dog’s breakfast of BTW and Openreach presumably passing the buck back and forth between one another. I also realise that there’s only so much AA can do. Nevertheless, I’m not happy at all. AA has offered to give me some help setting up L2TP should I need it.
When Janet and I are both feeling a little better (ie somewhat less ill both) I intend to proceed with my 4G investigations, which our very own kitizen meritez has helped me with. Later this week if we’re both up to it, I’d like to research the possibility of getting FTTP installed sooner rather than later, and without spending approx £105 to do it.
Does anyone know what generally happens regarding the politics behind the scenes when an ISP gets stuck trying to deal with BTW or Openreach regarding diagnosing or fixing a challenging line?
-
Some weeks back I got some rather upsetting emails from A & A about the problem I was having with my then HCD-inflicted line 2. Line 2 has since recovered on its own and is now perfect. Instead, line 4 has now caught extremely bad HCD and is almost permanently down now, but when it’s up it’s unusable, with so many errors and such a slow downstream rate that it needs to be disabled. In respect of line 2 AA said something that was a bit confusing but it sounded as if possibly AA had run into big trouble with either BTW and/or Openreach and one or the other was I assume refusing to fix the fault but I couldn’t make it out. Anyway, AA said they had reached the end of the road and couldn’t fix line 2. Today’s email, surprisingly, and disappointingly seems to extend this to line 4 and they’re refusing to fix line 4’s HCD either. Perhaps I’m supposed to be grateful that they haven’t just killed the line 4, or worse killed all of my DSL lines.
I realise that AA isn’t even remotely in control of this and are always reliant on this dog’s breakfast of BTW and Openreach presumably passing the buck back and forth between one another. I also realise that there’s only so much AA can do. Nevertheless, I’m not happy at all. AA has offered to give me some help setting up L2TP should I need it.
When Janet and I are both feeling a little better (ie somewhat less ill both) I intend to proceed with my 4G investigations, which our very own kitizen meritez has helped me with. Later this week if we’re both up to it, I’d like to research the possibility of getting FTTP installed sooner rather than later, and without spending approx £105 to do it.
Does anyone know what generally happens regarding the politics behind the scenes when an ISP gets stuck trying to deal with BTW or Openreach regarding diagnosing or fixing a challenging line?
Could you not quote exactly what was said please Weaver, rather than what you perceive it to have said .... especially if you're under the weather and maybe not seeing this as you would do, when you're at full strength ?? I think it has a massive bearing as to whether AA are refusing to accept your fault tickets, or BTw/OR.
You know me well enough by now and I wont bull-sh1t you mate ... but as I have mentioned time and again, there is a 'cone of acceptance' that all circuits have, and if L2 and L4 are showing as being in that cone, then it's always going to be a battle.
I will reiterate again for the trolls, I am on your side Weaver, I'm just trying to be realistic with your predicament.
I seem to recall there are other premises near to you, surely they too would be reporting issues as they are on the same extremely long cable run ??
-
I realise that I didn’t explain the level of severity, cone of acceptance thing, so let me rectify that. Line 2 is now perfect, that was the earlier killer problem referred to in the earlier emails from AA. It mysteriously fixed itself.
Now the current problem line, line 4, is either not syncing 95%, and that is over of the time, or very occasionally it comes up and syncs at 288-400kbps downstream with a ridiculous error rate. So it’s not remotely ‘marginal’ at all, the nice clearcult case that an engineer might like because there’s no doubt about it? Line 4 is effectively simply completely knackered: either dead, or worse. Note that I forced the downstream target SNRM up to 15 dB to try and keep the CRC / ES rate down to something tolerable, but even that SNRM that wasn’t enough to control the errors.
I’m now keeping line 4 disconnected so it can’t pollute the rest of the bonded set with all its CRC errors. Today AA said the only thing to do is to cease it and they didn’t offer to book an engineer.
Direct quotes from three earlier emails, as requested. Cut down for reasons of privacy and relevance / brevity. First email:
We have been trying for years to resolve this, and so far have failed to do so.
We don't admit defeat lightly, but I think it may be time to throw in the towel.
I think it's been clear for a while now that neither AAISP or BTW will be able to do any more than has already been done. This is of course not the outcome we wanted, and I'm sure neither did you, but we don't see a way forward.
Second email, excerpt:
[…]It's also worth mentioning, […], that any more fault reports will likely end in deadlock with BTW.[…]
Today’s email, which was extremely brief anyway, excerpt:
Line four is back up but in a very poor condition all we can suggest it to cease this line now.
A lot more detail would have been welcome in the first email, but maybe there are reasons why AA can only tell me so much.
A point about the other premises thing. Unfortunately I have no idea as I’m completely isolated. I never see anyone. I would have to ask Janet about neighbours’ experiences with copper lines. I don’t know how many neighbours use Openreach copper. It’s extremely dependent on which line you get. You either get a good line or a bad one. My line 3 is fantastic: 3.1 Mbps downstream sync rate and around 0.6 Mbps upstream.
Many recent HCD faults in the past were simply fixed by turning up and doing nothing other that running tests. Other faults were fixed by remaking joints, and this happened many many times. That seems to be the difference between a good line and a bad line: the joints you get.
For well over ten years with AA, or over fifteen years with DSL, my copper lines were reasonably reliable, with - I’m totally guessing, would have to many many review old posts - one to two faults per line per year. It looks like I get a lot of faults, but that happens when you have four times as many lines as everyone else. ;) Now for reasons unknown - was there a lightning strike ? - at the start of 2020, the storm of HCD faults began, and no one ever got to the bottom of it. The HCD faults were not confined to one line either.
So we had a good decade and a half and then three years of madness.
-
Here’s a picture of line 4’s performance over the last 20 hours or so taken from clueless.aa.net.uk, with the usual AA CQM graphing (https://support.aa.net.uk/CQM_Graphs):
(https://i.postimg.cc/SxKbmwTH/95-EFCEA2-90-EE-4-D73-AC05-841558-DEB5-DB.jpg)
Notice the ‘dripping blood’ in the graph.That’s what made line 4 truly unusable, even at a forced downstream target SNRM of ~15 dB. When line 4 was rarely up, I wished that it wasn’t, as it was ruining TCP performance. I assume that Zoom video conferencing wouldn’t work properly with uncorrected CRC like that, based on my speculative, dodgy theory that the likes of Zoom doesn’t use TCP or a similar retransmitting error-correcting transport layer protocol. (I have no idea whether or not that’s correct.)
-
It very much sounds like the 'bottom line' has now been breached, and it's no longer viable to provide you with a DSL service on L4.
Of course, it's all guess work on our behalf, as we're not involved with it personally. But I would imagine the amount of fault tickets raised against faults found are probably in the extreme ??
Not judging here, just (as always) pointing out the suite of tests that OR have to follow once a fault ticket has been raised ... you'll know them like the back of your hand now ... and the action to be taken based upon those results. If, the cone of acceptance (CoA) is achieved on all tests then ... you know the rest.
The tests have all been agreed with the ISP's. The tests are all logged. The tests can all be viewed by the ISP.
I've also mentioned a very powerful tool that has been available to engineers recently (and I would imagine the 'high repeat report triage team'), that can look at ALL broadband circuits and can be drilled down to single DP's (poles etc), or a number of same locality DP's, to check on how each circuit is performing in relation to the others.
When I say the engineer can go down to a granular level, I'm totally under-stating it ..... it is a fantastic tool for highlighting REIN and can even pull of a PSD mask of each circuit.
My own feelings are that there are probably a number of things coming together that dictate it is nobody's interests at all to maintain L4, financially of course.
I've seen it before probably 12yrs plus ago now, whereby I had a geezer who was like our very own Max on here ... there were engineers visiting site almost daily and nothing was ever found to be faulty. It took about a year, but eventually faults were refused to be taken against the gentleman, unless a definite line condition was found on the remote RAT test.
You may be a victim of your own scrutiny here, there were obviously times where a hard fault existed on the circuit, but there may have been many more where an intermittent burst of errors were also seeing a fault ticket being raised ??.
As I said above, I'm not involved in this in any way, shape or form ... I can only surmise this to be the case. I wish you all the luck with it, but more luck on getting FTTP. You are never going to have an error-free circuit on DSL where you are, mate.
-
Well that’s the thing. My line 3 for example is perfect, it is the error-free example. Quite literally, there are zero CRC errors either upstream or downstream and that’s even at 3 dB downstream target SNRM. So my lines are extremely good, thanks to the very very low noise environment because of the lack of human electrical interference near the cable, and because the copper is very thick, if I remember you correctly.
Just to clarify, I don’t ever call OR out trying to ‘perfect’ a line. If a line is a bit sub-par I ask AA what they think and then whether or not to do anything about it is up to them. So quite honestly, I never abuse the system trying to get the moon on a stick. With the recent rash of HCD faults, which are all extremely clear, I’ve even shown the HCD SNR-vs-tones graphs to engineers where I can, and the fault report is always because of ridiculously low downstream sync rate. I haven’t called OR in because the downstream rate is a bit slow; I’ve always waited until the problem is at its very worst, so that engineers can clearly see it, without any doubt and no cone of acceptance thing, because we’re down to (very roughly) 20% of the typical downstream sync rate say, and way way below the FTR. So there’s never been any doubt or disagreement about whether a line is ok or not.
So many times in the spate of recent HCD problems, though I can measure a problem, not just showing up in HCD depth but in appalling downstream low sync rate, and AA can see it too, but by the time an engineer appears, the problem has vanished and there’s nothing at all to see. This may or may not be due to something mysterious that the engineer has done themselves inadvertently. That is the very great mystery of the past few years. Yet on the other hand there have been very many instances where the engineer has found the cause of the HCD problem, usually some bad joint and has mended it quickly, with no trouble.
-
. . . the copper is very thick, if I remember you correctly.
0.9 mm diameter copper (cores of the) pairs is my recollection.
-
‘Tis mine too.
Anyway, as my parents would say, “I’m a bit done in”.
-
It is indeed the 0.9mm Cu feeding your premises, well remembered.
My sole aim here, is trying to understand why the refusal to fault one of your lines has been agreed ??
There has to be validity to it, as I suppose you could flag this up to Ofcom and they (knowing this) would have to provide some kind of supporting evidence as to why services have been withdrawn ??
They can't simply turn round and say 'He's a nuisance' :) :)
Just to cover old ground again here, probably 5'ish years ago, a 'High Repeat Report' team were established to cover exactly your kind of issues.
Having been involved with broadband from the very start, I can tell you that in the first few years the training was woeful and the engineering support was too. I myself hold this very forum on a pedestal, as a massive aid to my own personal development in this arena.
So, please believe me when I say that the 'HRR' team that was set-up, was manned by very knowledgeable individuals who could scrutinise circuits along with the best. Coupled with access to other systems mentioned earlier in this thread ... they were found to be a fantastic asset as the time consuming hard-work had all been done, by the time the engineer had picked the fault up.
What I'm trying to imply I suppose is my guess is that somebody of that ilk has done a whole lot of cross-referencing of previous fault reports, outcomes of said reports, general stability of the circuit, performance in relation to other nearby circuits etc etc ...... and decided to draw a line under it.
Not for one minute am I suggesting you don't have an actual fault condition on L$ (or whichever one it is ??), I'm trying (and maybe failing) to get the message over to any of the readership viewing this, that over-scrutiny and constant fault reporting can have an adverse effect if the perceived issue falls within the ISP agreed 'Cone of Acceptance'.
I reiterate, there has to be a valid reason as to why they have stopped raising fault tickets.
-
I still get an eye twitch every time you say "Cone of Acceptance" though, as how can any line failing to sync for hours at a time, and syncing with too many errors to be usable, be considered within it?
-
Reading your posted quotes it sounds to me that A&A are throwing in the towel.
They are assuming that any more fault reports to OR/BTW "might" end in deadlock.
That should be your choice to make if you want to push the issue that far or they should be honest and say we personally are done with this line.
-
I still get an eye twitch every time you say "Cone of Acceptance" though, as how can any line failing to sync for hours at a time, and syncing with too many errors to be usable, be considered within it?
It doesn't - like I have said above, it could be the other shed-loads of faults that have been raised that were probably within the CoA, or the tweaking of the DLM down to 3/4/5 dB, or whatever ........ that have all played a part with the end result being that on the whole, the circuits MPF is perfectly fine.
I've said it three times now - there HAS to be a reason why they've pulled the plug. Whether Weaver will ever get the real reason or not, who knows. But it reads to an outsider looking in, like AA/BTw/OR have got enough justification to do as they are.
-
It doesn't - like I have said above, it could be the other shed-loads of faults that have been raised that were probably within the CoA, or the tweaking of the DLM down to 3/4/5 dB, or whatever ........ that have all played a part with the end result being that on the whole, the circuits MPF is perfectly fine.
I've said it three times now - there HAS to be a reason why they've pulled the plug. Whether Weaver will ever get the real reason or not, who knows. But it reads to an outsider looking in, like AA/BTw/OR have got enough justification to do as they are.
And, suppose that *is* the case - that is really poor of whoever is to blame (I'm not going to call any specific party out on that one) - especially as Weaver, as I am sure he won't mind my saying - depends on decent internet and he isn't in a position to easily go find an alternative. Given how much an AA line is each, £35 and Weaver has, 3 now, I think? Not a small amout of money..
-
And, suppose that *is* the case - that is really poor of whoever is to blame (I'm not going to call any specific party out on that one) - especially as Weaver, as I am sure he won't mind my saying - depends on decent internet and he isn't in a position to easily go find an alternative. Given how much an AA line is each, £35 and Weaver has, 3 now, I think? Not a small amout of money..
It's a tricky one for sure, as Weaver's been around this forum for quite a while and we all know his situation.
Trying to see the debate from all sides (as everyone should in any conversation), the 'suppliers' are running a business, they shouldn't be expected to run at a loss .... and it certainly looks to me at least ... like one or all parties involved, have decided enough's enough.
Again, I feel Weaver may have been a victim of his own scrutiny/tweaking .... we don't know the facts, probably never will ... just people's opinions. But I would love to see if Weavers immediate neighbours circuits are batting along just fine, with minimal/no fault reports raised against them ??
Most folk just want the circuit to work, they won't have a clue that the DLM has nicked a bit of speed due to x,y,z reasons. They certainly won't be monitoring their error counts and laying with the DLM settings.
That is where the full picture would show ... if the neighbours are all content but Weaver is generating tens upon tens of fault reports, then it will eventually get looked into and this is where we are today.
You have to take sentiment out of this, harsh as that sounds, but it isn't AA Charity Comms, it isn't Openreach Gift Aid .... they are businesses. I'll say it a fourth time, there has to be some amount of data that one or the other (or both) have compiled to come to the conclusion that the MPF is good, ie: within the CoA.
I don't make the rules, Weaver doesn't make the rules .... OfCom, the ISP's and OR do.
-
Oh, and I hope I wasn't coming across as being critical of any one business. I totally understand where you are coming from on that one and am in total agreement with you. Ultimately - we are stuck with copper for a bit longer - and really we pushing the boundaries of what we can do when you consider the distances involved, plus the remote location where Weaver is.
It was never designed to send internet connectivity at high speeds over long distances...
-
No, no - absolutely you weren't.
Likewise, my own text may be being read in the style of some heartless & pompous you-know-what. Not the case at all, I genuinely feel for Weavers predicament but am also trying to see it from the other side of the fence, too.
No point us all patting the poor geezer on the head, saying how monstrous they're all being. That time might be better spent trying to understand why they've come to their decision .... IMHO, of course. ;) :) :)
-
To reiterate what I said before, it has always been AA’s choice not mine as to whether or not they get on to BTW/OR. And if you want to know about other lines then you can ask me. My lines 1 and 3 have always been fine, in fact line 3 is perfect. The third point is that sometimes even with HCD faults, OR engineers have found a fault out on the moor and fixed it by repairing a joint. So it’s not the case that there are non-problems reported. There most certainly have been problems that have vanished between the time I report them and the time when OR arrives but I can’t help that and in every case I have objective proof of the extent of the HCD fault and as I have said before have always left it to get incredibly bad before calling for help, so that an engineer will have something obvious and really bad to see. There was a fault that was a no-show when the OR engineer arrived, no problem seen, and that could have been because AA understandably called OR out too early, trying to get it fixed quickly, but it was too soon and hadn’t developed as much as I would have liked to give OR the best chance. But once again, AA makes the decisions as to whether or not to call OR out, not me.
I do not order AA to fix non-problems and you can believe that or not, unfortunately I can’t easily prove it. But you could look back at all the old threads.
As for the ‘problem of scrutiny’ I’ll just have to ask you to believe me that I did not ever do this. I’m sufficiently clued up to know what is a problem and what is not, and since anything bad gets handed over to AA it is they who make the decisions not me. My emails to AA always contain a ‘what do you think?’ or ‘would you take a look at x?’. I don’t nag AA into fixing faults, or non-faults for that matter. You don’t really know me, and I could be one of the OR time wasters that you have encountered and like certain persons I can think of, but I just ask you to believe that I’m not and in any case, there is always AA there who would act as a filter.
I haven’t tweaked anything to run at 3 dB SNRM downstream. It’s an option on AA’s control panel, and I discussed it with AA first, and monitored the results to see if there were any errors. If there had been any downstream errors then I would have put the SNRM up to 6 dB. I’ve been very happily running at 3 dB downstream SNRM for well over five years and the only reason that I can do it is because of my Broadcom PhyR software in the modems and that in the exchange which fortunately matches it. It’s easily worth about 4 dB. Unfortunately it’s only available for downstream.
AA is understandably saying that they simply can’t diagnose HCD after trying for many many months.
Recap: ultimately this leads to very high downstream SNR and extremely low downstream sync rates, say 288 kbps - 400 kbps instead of 2.0 - 3.1 Mbps often with high packet loss too. In anyone’s book those sync rates are unusable, no?
We’re not talking about a minor whinge or a request for optimisation. Occasionally bad joints have been found, fixed and that cured it. So it’s possible that all HCD is caused by bad joints.
-
OK Weaver - what is your own personal opinon of why faulting service has been withdrawn ?? Do you think it's AA, or OR ??
I'd humbly suggest from what you've said about never pushing the issue yourself, that it is AA that have taken the stance to stop service ??
-
I do not order AA to fix non-problems and you can believe that or not, unfortunately I can’t easily prove it. But you could look back at all the old threads.
As for the ‘problem of scrutiny’ I’ll just have to ask you to believe me that I did not ever do this. I’m sufficiently clued up to know what is a problem and what is not, and since anything bad gets handed over to AA it is they who make the decisions not me. My emails to AA always contain a ‘what do you think?’ or ‘would you take a look at x?’. I don’t nag AA into fixing faults, or non-faults for that matter. You don’t really know me, and I could be one of the OR time wasters that you have encountered and like certain persons I can think of, but I just ask you to believe that I’m not and in any case, there is always AA there who would act as a filter.
I believe you, for what it's worth. But I can certainly see that for an outfit like AA who basically trade off of their reputation for excellent support and fault-fixing prowess (and yes, they charge appropriately for that too) they may feel the need to push for resolution of a more marginal fault even if the fault ticket was only opened to get their view on something.
I suspect if your ISP was a bog-standard mass market ISP there would have been a lot fewer faults raised with Openreach even if you raised the same number of faults with the ISP, simply because they would have been happier to say that some of the more marginal errors are just due to line length and there's nothing that can be done, than AA will have been.
-
I’m just completely confused and haven’t been able to form an opinion. There was the mention of deadlock so that might indicate that there has been some barney between AA, BTW or OR. Or it’s just as likely that this is coming from AA, and thinking on that side, I’m quite with you.
Because it has all been so vague, I asked them straight out, what will happen if I do what I’ve done before successfully, order an additional line and then cease the bad one and they said explicitly that I was allowed to order a new line according to the plan. I was given quite proper and helpful warnings that the plan probably wouldn’t work. Later on I said to them that I had taken heed of their advice and decided not to go ahead with the plan, ie not to order an additional line.
Mind you, I’m starting to change my mind, thinking about ordering a new line, my reasoning being what’s the worst that can happen - well I end up with a new bad line and have wasted a bit of money, but apart from that I’m not any worse off.
Actually now I stop to think about it the worst that can happen is that I order a new line, it all looks good and then AA completely refuses to support it, that’s probably ‘the worst’. And paying for AA with support refused from the start, well that’s quite crazy.
I’ve been a loyal customer for well over 30 line-years and have paid AA a decent amount of money, so you can understand why I might not be too happy.
-
For the record I didn’t think or have ever thought you were pushy or demanding.
I always assumed AA took this on for you in good faith because they seem the sort that would care enough when you make them 3 x the business and they sell themselves as sorting out OR with no BS.
-
Thanks dee.jay, you’re the best.
-
Thanks dee.jay, you’re the best.
Haha no problem.
Just wish I could do more for you.
-
This is what I don't understand from what BS was saying, I never got the impression you were nit-picking little issues. You are paying AAISP for a bonded service and that inherently means that issues which would be minor for someone with a single line DO need dealing with, or that line is useless. That's not not-picking or overly scrutinising your lines IMO, you only brought up issues when that actually impacted your use of the service.
-
Looking at this in a slightly different way, more mathematically. If you have 4 lines into your property ( I think its 4?), and the MTBF is consistent with similar lines to neighbours properties, in total you are going to have 4 times the amount of faults to your property as a neighbour with only one line. ie having more lines increases the probability of a fault to the property.
-
@craigski - quite so.
Irrationally, I might feel an abuser just because I’ve got more lines and hence more faults, as you say. And just because one of my lines has a fault doesn’t mean that it is any more likely that my others will be faulty, not in this case, because faults can be fixed and I’ve never had any RF interference related problems which - you would expect - would affect all if they affect one.
I need to get on and try sorting out the 4G kit that meritez has helped me with. My health is better now, finally after months, triggered by me catching the ordinary for fir the first time in recorded memory. Although I never see anyone, Janet caught it off someone and gave it to me. I just need Janet to be available to help me with it, which is difficult to arrange.
-
@Weaver, I really do think you'd be best going the 4G and/or satellite route as it's starting to look like your landlines are becoming unsustainable from a DSL perspective. Openreach only have a certain level of obligation when it comes to providing an internet service and if you have access to 4G I'm not sure they have any obligation to provide at all - https://www.bt.com/broadband/USO
-
Been honest Weaver I am amazed you have managed to get all the call outs you have done so far. Your line is very long so I expect Openreach will have the point of view its just the nature of the service, and that you run them on an aggressively low 3db margin as well.
Openreach I expect are looking to the future and see copper as dead tech now.
AAISP been stuck in the middle are probably trying to do the best by you. However I dont think you been unreasonable either, they do sell themselves as dee jay said as been a master of dealing with BT line faults, I did indeed even post in my own thread not long ago I was disappointed with a response I got back from them on single threaded speeds which I felt was a mass market type response.
Did you look into your 4G L2TP idea? I think thats your stop gap solution until the FTTP is done in your area. More and more people use 4G at home now, and you can keep your static IP addressing.
Going away from DSL felt risky for me as I have been used to it for so long, but I absolutely do not regret it and L2TP allows you to keep your AAISP network stuff. Also would be no more stress of monitoring fragile DSL lines, DLM etc. things are more of a "just works".
--edit
I have seen some of the points I mentioned have been raised in the thread, my suggestion Weaver is to treat the 4G migration as a priority, I think it will be something you wont regret. The DSL problems you have, have just been dragging on. If needed I dont mind in DMs giving you my experience of setting up my L2TP, the MTU etc.
-
Well 3 dB doesn’t ‘break the lines’, despite what some superstitious folk have said to me. I run at 3dB downstream where I can do so with zero ES. If I do get any errors, then I increase to 6 dB. And that has nothing at all to do with the bad joint faults that have been reported and fixed. I have absolutely zero sympathy for Openreach as the knowledge of the engineers who have come out to me has been truly shocking, but they can’t find faults that aren’t there. The faults have gone into hiding before the engineer arrives or else some mysterious action by the engineer has fixed them. I don’t understand why Openreach doesn’t escalate the issues given that there have been so many repeat faults mended again and again. Look back in the old posts, everything is documented. And as for the ‘copper is dead technology - don’t support it’ philosophy, that’s pretty callous - what they should be doing is replacing the worst copper with FTTP first, which they have done in some extremely remote parts of Skye. We’re still BT customers and they aren’t allowed to just forget us. It is a complete waste of time re-fixing the same faults again and again, so why don’t they fix it properly, for good or far far better get rid of the copper right now. `they must know how bad some of the lines are in reliability terms.
I have already been looking into 4G hardware with Meritez’s help, but illness over the past few months has meant endless delay. Since my own health is now picking up and so is that of my wife hopefully, maybe we will get something done before xmas.
AA has been just totally confused by HCD. I suspect they’ve never seen it before. No kitizens have come back saying it’s familiar. And I wouldn’t be surprised if BT has not issued all kinds of threats to AA - complete speculation, based on nothing, and we’ll never know,
Offers of help greatly appreciated. AA has also offered to help me set up L2TP in the Firebrick. But I’m nowhere near that stage yet. I haven’t even managed to gain access to the modem-router yet, that’s been my awkward problem. It all depends on health or rather lack of it.
I wonder if I can bond DSL/PPPoE with L2TP/4G ? Would MTU problems make that a nightmare, or do I just reduce the MTU to the lowest common value?
-
I just hope you have better luck with L2TP than I do, mine went down for a while recently because pfSense has a nasy habit of not reconnecting when there is a momentary glitch with 5G and I get fed up of manually forcing it. Its supposed to be my backup into the home network if I'm ever away from home and the main connection goes down, kinda useless if it can't be relied upon to stay connected.
Mobile is great for speed, but I'm very dubious about using it for reliability as I'm in an area with plenty of cell density, decent broadband, and its still not exactly reliable. Somewhere remote where there will be few cells and people hammering it, I'm just not sure its going to be an improvement. Hopefully the Firebricks down have to take down the main connection every time the backup goes wonky, pfSense does and it sucks. I've had more downtime due to 5G dropping causing the main connection to briefly go offline, than I ever had with the main connection dropping itself.
-
I don’t understand why Openreach doesn’t escalate the issues given that there have been so many repeat faults mended again and again.
They do - it's called the FVR (Fault Volume Reduction) programme.
If a certain area has a high amount of repeat fault reports, then the powers that be will look at the historic cost of 'truck rolls' per fault (used to be around £350) and make an educated decision if an upgrade would be frugal.
Which also leads me to believe, you are the only individual reporting faults in your locality.
-
Weaver you can just reduce MTU on the DSL if you want it to match. Although you might find once you getting circa 50+ mbit on the 4G, you wont need to bond for bandwidth, that will feel like a huge amount given what you been used to, so might be better to just configure the DSL as auto failover.
Alex, mine has been ok on pfSense, but on windows, when AAISP had an outage, it just disconnected and stayed off, pfSense if that happens reconnects fine, although I have increased the fail parameters on the monitoring to make it more lenient, as I would rather my policy routing not switch for a very short drop.
-
They do - it's called the FVR (Fault Volume Reduction) programme.
If a certain area has a high amount of repeat fault reports, then the powers that be will look at the historic cost of 'truck rolls' per fault (used to be around £350) and make an educated decision if an upgrade would be frugal.
Which also leads me to believe, you are the only individual reporting faults in your locality.
Interesting, when my last engineer visited last month I learnt something about work been done in my area.
Some may remember my DSL jumped up from circa 60 mbit to a fairly consistent high 70s, 80mbit, I assumed it was from when they fixed the tie pairs at the cabinet.
Well it turns out when I told the engineer my line was syncing way higher than the line estimates, he said the estimates are based on old data and in the area they actually did some copper rerouting, he said it was unusual but was a lot of speed faults in the area at the time. Even though he couldnt fix anything on the call out we had a very good chat.
-
Alex, mine has been ok on pfSense, but on windows, when AAISP had an outage, it just disconnected and stayed off, pfSense if that happens reconnects fine, although I have increased the fail parameters on the monitoring to make it more lenient, as I would rather my policy routing not switch for a very short drop.
The problem is its regional, if your tower gets periodically overloaded then no matter what parameters you end up with the firewall restarting constantly and killing connectivity. So I had to turn off monitoring entirely as having your backup connection cause your main connection to be unstable defeats the point of having one.
It doesn't help I've had the 5G router go offline a few times needing data toggled to get it back.
-
Well it sounds like he isnt cancelling the DSL, so if it doesnt work out he is just back where it was, configure the DSL as backup.
I do sympathise with Weaver, he is paying a ton for multiple business lines with higher priority faults as well, so he is paying a fair whack for his service.
Also 5G has problems that 4G doesnt have, I deliberately disable 5G on my phone.
There is people who use 4G as their day to day broadband, and every time I have used it, it doesnt drop out, has been reliable, so its worth him trying.
On your connectivity going down on firewall restarts, if you using pfSense 2.6.0 enable the PF counters patch, a new feature causes massive latency spikes on every firewall reload, the patch disables the feature. To be absolutely clear I have never had a 4g disconnect when hooked up to pfSense. Its actually DSL isp's that give me uptime headaches.
-
I've used 5G and 4G, and while 5G has problems more often (likely as its adding more traffic to the 4K upstream until they roll out of proper 5G), they both had days where it was practically unusable and were talking completely different networks/cells.
Not sure what 5G specific issues you mean other than the above, given 5G is supposed to improve congestion issues over 4G.
-
I've used 5G and 4G, and while 5G has problems more often (likely as its adding more traffic to the 4K upstream until they roll out of proper 5G), they both had days where it was practically unusable and were talking completely different networks/cells.
Not sure what 5G specific issues you mean other than the above, given 5G is supposed to improve congestion issues over 4G.
Well what are you trying to say here, are you suggesting Weaver should not try it because you think 7km ADSL lines are better that the ISP has said is now a wont fix?
5G specific issues related to very spotty coverage, its coverage is a fraction of 4G and the radius is much smaller as well with lower indoor penetration.
As an example my phone gets a only a few hundred kbit on 5G, but 24/7 120mbit on 4G. 5G needs lots of extra masts to work as well.
-
Well what are you trying to say here, are you suggesting Weaver should not try it because you think 7km ADSL lines are better that the ISP has said is now a wont fix?
No I was saying whatever he does DON'T get rid of the DSL and don't be surprised if his Zoom calls (or whatever it was he was doing) still suck. ;)
I guess I shouldn't worry given his networking knowledge and he is already familiar with 3G backup.
-
No I was saying whatever he does DON'T get rid of the DSL. ;)
Regardless, I suggest you enable that pfSense patch as your connectivity going down on a firewall reload isnt normal. (unless you enabled the option to reset states).
-
If there's one thing I dislike about Netgate its how long it is between releases. They seem to be grossly understaffed for bugfixing.
-
If there's one thing I dislike about Netgate its how long it is between releases. They seem to be grossly understaffed for bugfixing.
There is a fair amount of niggly issues yeah, they will fix promptly if a patch is provided, but when its just a bug report it can be left at there for years, I think they usually prioritise things that affect the paying customers.
I have about 20 patches I made running on my unit right now, some of them will be submitted but some wont be as I think they wouldnt accept, I do plan to put them on the kitz wiki page soon.
-
I'm always a bit worried patches will break future upgrades, or are they applied on-the-fly so the installation remains stock?
-
I'm always a bit worried patches will break future upgrades, or are they applied on-the-fly so the installation remains stock?
Those patches in the screenshot are done by the pfsense devs who maintain the package, These dev supplied patches I think will only be allowed on the right specific pfsense build (note it says 2.6.0 in the black box), these dev supplied patches I believe will all be part of 2.7.0, they just made the patches to get the fixes out quicker.
Above those patches which I didnt screenshot is my custom patches.
You can choose to have them auto reapply on pfSense updates or to leave auto update off so if the code changes, the patch will be there available but not applied.
However there is sanity checks, it wont apply the patch if it detects its not compatible (like if only some lines would patch instead of all). It will hide the apply button, and auto apply will be overruled if its not a clean patch, I used to auto apply, but now I dont, and just put 'ON' at the end of the name to signify I currently use that patch so I know which ones to reapply after an update. In short if you disable auto apply, any updates will remain stock until you decide to apply the patch, patches applied can be easily unapplied by clicking revert.
-
Weaver I hope the little back and fro between me and Alex didnt put you off, please let us know how you get on and what you decide.
I assume if you do get wireless, it will be to supplement and not replace the DSL as Alex suggested.
-
I’m going to ask AA if I can bond DSL and L2TP, provided I sort out MTUs properly.
-
I’m going to ask AA if I can bond DSL and L2TP, provided I sort out MTUs properly.
If 4G can give you a decent service I'd have thought you'd be better using that and having perhaps one DSL line as an emergency fallback?
-
I would agree surely a good 4G will be better service than 3x ADSL lines so bonding it seems worthless. The DSL backup would be a good solution. No need for expensive kit to do this.
I would suspect @weaver might not like the stats however from a 4G service !
-
I suspected the bonded DSL will still be needed for any live video conferencing, normal traffic will probably be okay over 4G.
I mean in all honestly I did GAME over 4G and 5G which normally was fine, but its those days/weeks where it suddenly wasn't that were an issue. Especially in an area where 4G is the only fast connection, I'd be surprised if it performed well at peak hours and I was literally seeing ping spikes to 2000ms on those bad days (may be inflated by ping traffic being lowered in priority due to the contention).
Bonding over 4G when it has one of those bad hours/days/weeks would be a disaster. IMO policy routing latency sensitive traffic over DSL is a must. The way I currently use 5G is as failover and load-balanced for bandwidth heavy traffic, such as Steam download servers. I originally was going to use it for video streaming too but I wasn't getting consistently good bitrates on that.
-
I would agree surely a good 4G will be better service than 3x ADSL lines so bonding it seems worthless. The DSL backup would be a good solution. No need for expensive kit to do this.
I would suspect @weaver might not like the stats however from a 4G service !
Would solve all his packet ordering problems.
-
I don’t know how much packet ordering misery I may be getting. It doesn’t seem any worse with more lines than fewer and the downstream speed is superb in terms of n times the expected rate. Upstream speed is normally really rubbish in terms of n times rate but in the past, as I have noted in old threads, there have been periods when it was excellent. Goodness knows how or why. You would the the FB2900 has the same code for packet scheduling upstream as the AA 6000/9000 series routers do for downstream; the latter are great, the FB2900 may possibly be rubbish in this respect. I would think that good design if the router could avoid packet reordering but if it occurs further down the path then there’s only so much that you can do about it without dodgy compensatory tweaks that could make things slightly slow in certain scenarios. By that I mean scheduling a ‘later’ packet in the flow to be really late, so that it will arrive suitably late, rather than sending it up link 2 in parallel straight away. It should take notice of the packet lengths at least and use that in the ordering, but perhaps a lateness tweak configurable parameter should be an option.
-
We're in a somewhat similar position, AA and BT reached deadlock on our DSL with BT declining any further action on high error rates or PPP lockups on the basis that these are supposedly caused by an electric fence. That's an electric fence alongside a 50 pair cable, which apparently affects our line but nobody else. Honestly I was quite wound up about this at the time, but I am now over it.
On the plus side BT have in fact cleaned up the line, knocked a couple of dB off the attenuation and increased synch speed by around 15-20%. AA on their part have completely disabled DLM so an error rate of 2800 ES/hour doesn't actually slow the line down at all. The lock ups are still a real nuisance, sometimes only once every few days but sometimes constantly for hours on end.
The solution for me is 4G, and that works perfectly well for video conferencing as I've found working from home since Spring 2020. As a matter of fact Webex doesn't use that much bandwidth, MS Teams is more needy and really needs more than ADSL upload in any case. I'm using Three on Band 20 giving around 20meg down and 10 up.
The only 4G issues I experienced were with Smarty where some services including my employer's CRM objected when we started getting second by second IP address changes. Smarty never owned up to any changes in their CGNAT, but moving to Three with a real IP put that problem behind me.
-
Wow, thanks for that recap. I suspect that that puts Smarty out of the running. A normal Three data SIM that is unlimited usage or practically so is something I can ask Meritez about, our deals and 4G guru.
-
Yes indeed. One thing I forgot to mention is that until about a year ago Three's coverage checker said we couldn't get 4G at our address. I tested with free PAYG SIMs from Three. Even without putting any credit onto them they give 200meg, topped up every month so each SIM good for a few tests. On the back of those speed tests I went ahead with Smarty that's a month by month commitment. Initially I held off taking a committed contract for a service that Three themselves said we couldn't receive. I was on Smarty from March 2019 until sometime in 2021 when I jumped ship to Three direct.
ADSL is currently retained and configured as backup using Mikrotik's "recursive routing". This configuration allows me to selectively configure some traffic to prefer DSL, with failover to 4G should I so wish. I don't pass through, the WAN IP address terminates on the SXT which does NAT and firewall for 4G.
This has all reminded me that I keep meaning to post for comment about my failover configuration and topology, but I'll really need to create a drawing before I do that.
-
Wow, thanks for that recap. I suspect that that puts Smarty out of the running. A normal Three data SIM that is unlimited usage or practically so is something I can ask Meritez about, our deals and 4G guru.
@Weaver, you would be looking at a 24 month commitment at £12 a month for a three business unlimited data SIM.
https://www.scancom.co.uk/pages/search-results-page?q=december&tab=products&sort_by=price&sort_order=asc&page=1
You may be better on EE depending on speed tests carried out locally.
-
We're in a somewhat similar position, AA and BT reached deadlock on our DSL with BT declining any further action on high error rates or PPP lockups on the basis that these are supposedly caused by an electric fence. That's an electric fence alongside a 50 pair cable, which apparently affects our line but nobody else. Honestly I was quite wound up about this at the time, but I am now over it.
On the plus side BT have in fact cleaned up the line, knocked a couple of dB off the attenuation and increased synch speed by around 15-20%. AA on their part have completely disabled DLM so an error rate of 2800 ES/hour doesn't actually slow the line down at all. The lock ups are still a real nuisance, sometimes only once every few days but sometimes constantly for hours on end.
The solution for me is 4G, and that works perfectly well for video conferencing as I've found working from home since Spring 2020. As a matter of fact Webex doesn't use that much bandwidth, MS Teams is more needy and really needs more than ADSL upload in any case. I'm using Three on Band 20 giving around 20meg down and 10 up.
The only 4G issues I experienced were with Smarty where some services including my employer's CRM objected when we started getting second by second IP address changes. Smarty never owned up to any changes in their CGNAT, but moving to Three with a real IP put that problem behind me.
How did AAISP disable DLM I thought it was a Openreach thing? Also glad you confirmed how well 4G can work as I felt I was really the only one in this thread who could vouch for how good the tech has got. No issues with video calls, browsing, streaming (download streaming as well as me streaming out to the internet), downloads (including single threaded as can see in my VDSL single threaded issues post) and VOIP.
To weaver if Three 4G isnt great, remember you can try another provider like EE as an alternative. EE probably have the best 4G in the country still.
-
AA can disable DLM for me and AESmith because we’re mere ADSL users, not FTTC.
-
Ok thanks for that, so the ISP has been given control for ADSL, thats interesting.
-
I didn’t know that at all until AA disabled DLM on one of my lines too. That has its downsides, I have to say. Having been frustrated with it when it was here, I am perhaps almost missing it now DLM is gone on that line. I can’t believe I just said that.
-
To weaver if Three 4G isnt great, remember you can try another provider like EE as an alternative. EE probably have the best 4G in the country still.
As far as I know Three are the only provider who give a real IP address, rather than CGNAT. I'm referring to IPv4 though.
-
As far as I know Three are the only provider who give a real IP address, rather than CGNAT. I'm referring to IPv4 though.
Thats true I couldnt find a way to not have CGNAT on EE, but if he uses L2TP would it matter?
-
I guess not, but the AA L2TP adds £10/m. If using your existing AS account, L2TP takes precedence, if it's already connected then DSL won't authenticate. If DSL is already connected then L2TP authenticates but doesn't blow DSL off, and routing gets screwed up. That's what I found anyway.
I guess Weaver will want IPv6 though.
-
As far as I know Three are the only provider who give a real IP address, rather than CGNAT. I'm referring to IPv4 though.
EE provide a real IPv6 address on their mobile service.
There's nothing to stop you forcing EE to IPv6 only, and using the A&A l2tp over it.
-
EE provide a real IPv6 address on their mobile service.
There's nothing to stop you forcing EE to IPv6 only, and using the A&A l2tp over it.
Would that not be less efficient than over IPv4 due to the IPv6 having more overhead?
-
Ok thanks for that, so the ISP has been given control for ADSL, thats interesting.
It has been like that a number of years. It very much depends on the backhaul carrier.
BTW run their own DLM on their ADSL DSLAM's and Talktalk run their own on their DSLAM's.
I think Sky do too.
With Talktalk for years I had their ADSL DLM set to 3dB then disabled so it could never change from that. In roughly 2014/15 Talktalk changed their DLM so you could no longer disable it. They added that back some time later.
I haven't kept up with any of the ADSL DLM's since then.
I would assume the control A&A give on ADSL lines will depend on whether it is BTW or TTB.
Edit: looks like it does indeed.
https://support.aa.net.uk/BT_ADSL_Line_Profiles
https://support.aa.net.uk/TalkTalk_Wholesale_Line_Profiles
Looks like you can't set 3dB
-
Would that not be less efficient than over IPv4 due to the IPv6 having more overhead?
I'm all ears, what overhead?
https://www.ipv6.org.uk/wp-content/uploads/2018/11/Nick-Heatley_BT_EE_464xlat_UKv6Council_20180925.pdf
https://indico.uknof.org.uk/event/38/contributions/489/attachments/612/736/Nick_Heatley_EE_IPv6_UKNOF_20170119.pdf
-
I’m sorry, I got lost. Is CGNAT a show-stopper for L2TP ? Does L2TP care about CGNAT addresses or changing addresses ? I can always ask AA of course. Selection criteria for choosing suitable 4G carriers compatible with the use of L2TP.
-
@Weaver, CGNAT is not a show stopper for L2TP, in fact most use L2TP to get round CGNAT.
-
Thanks, I assumed as much. Does it kill L2TP if source addresses change ? I will be pinging with ICMP pings every second or so to check that the connection is healthy, so that should perhaps help to keep a 4G link up, prevent something from timing out and dumping a source address somewhere.
-
AA's L2TP worked fine over Smarty CGNAT.
-
Thanks my friends, for clearing that up. So my concerns will be packet loss and worst-case speeds down and up. I’m sure I’ve forgotten something important in that list. The L2TP connection had better stay up long term, and be highly reliable.
In the other thread I mentioned that I’m now completely confused about what the protocol stack is coming out of my Firebrick into the 4G radio and then the stack over the 4G link. The wikipedia article about L2TP is no help.
-
Hitting the 4G will be IP over PPP over L2TP over UDP over IP over Ethernet. The Ethernet will be stripped by the 4G modem.
The 4G link itself will have an IP MTU of at most 1500 bytes with the rest not your problem: further overheads will be dealt with by the 4G network.
-
I’m sorry, I got lost. Is CGNAT a show-stopper for L2TP ? Does L2TP care about CGNAT addresses or changing addresses ? I can always ask AA of course. Selection criteria for choosing suitable 4G carriers compatible with the use of L2TP.
It shouldnt be an issue as its an outbound connection on the client. the L2TP server is AA's side, however even if it is, you can connect on L2TP via IPv6 I checked on the L2TP AA page after meritez posted. I also tested it successfully on EE, I think I mentioned that in my L2TP thread.
On reliability I have had a few outages which I assume are all AA related, one of them Andrew was woken up during the night and was definitely AA side, thats the only one that didnt instantly reconnect, all the others reconnected automatically. Its going to be a "try and see" thing and hope it works well. If it does then you can decide on longer term configuration changes.
You will still have control over the rate limit as well if you feel you want the small packets optimization, 100% rate limit is 200mbit.
-
I’m going to continue this in the other thread (https://forum.kitz.co.uk/index.php/topic,27208.msg461621.html#msg461621), so please follow me.