Line problems since BT work carried out last night.
About a month ago I was getting very slow speeds. At first I thought it was related to the BGN issues as it started at the same time.
However.. DaveT produced a graph showing that my SVLAN appeared to be running hot. Over the past few weeks speeds continued to decline and I'd see 20Mbps in the evenings, but I left it thinking BT would fix things soon. link
This morning at 04:18 my line lost sync and the PTM pathway was unavailable for an hour. My line came up been its been carp ever since - my upstream SNRm is normally rock steady, but I'll attach some graphs so you can see whats been happening since. My errors are going through the roof and no doubt the DLM will penalise me sometime soon.
This behaviour is not normal, I have a full 80/20 that is normally nice and stable with few errors.. so as you can see the graphs are pretty damn dismal. Ive done a full reboot of my router this evening to see if it helps but no.
The line seems quiet but something is generating all those errors. If I raise a report, then MrBTOpenreach is going to come and say - hey your line is performing within parameters... what do you reckon my chances of getting this sorted are?
Whatever BT did last night, it seriously messed with my line
btw... my speeds are fine tonight and no more 20Mbps.
----
PS looks like I lost sync at 20:45 - that wasnt me.
However.. DaveT produced a graph showing that my SVLAN appeared to be running hot. Over the past few weeks speeds continued to decline and I'd see 20Mbps in the evenings, but I left it thinking BT would fix things soon. link
This morning at 04:18 my line lost sync and the PTM pathway was unavailable for an hour. My line came up been its been carp ever since - my upstream SNRm is normally rock steady, but I'll attach some graphs so you can see whats been happening since. My errors are going through the roof and no doubt the DLM will penalise me sometime soon.
This behaviour is not normal, I have a full 80/20 that is normally nice and stable with few errors.. so as you can see the graphs are pretty damn dismal. Ive done a full reboot of my router this evening to see if it helps but no.
The line seems quiet but something is generating all those errors. If I raise a report, then MrBTOpenreach is going to come and say - hey your line is performing within parameters... what do you reckon my chances of getting this sorted are?
Whatever BT did last night, it seriously messed with my line
btw... my speeds are fine tonight and no more 20Mbps.
----
PS looks like I lost sync at 20:45 - that wasnt me.
The DLM clobbered me last night - Other than when I was testing something, this is the first time in >10yrs that Ive been interleaved.
Despite Interleaving and INP being applied, this has done nothing to arrest the errors and they still continue at the same rate.
I still have the same amount of CRCs and ErrSecs.. I just also now have lots of FECs too... so interleaving has done naff all other than to increase my latency by 7ms.
Ive swapped my router back to the HG612 but that makes no difference either.
I think Im going to have to raise a ticket.. but because the error rate has continued, Im unsure if the DLM will get me again tomorrow.
Im still also loathe to raise it because right now I could do without BT hassle and them saying there's nothing wrong with your line - look it syncs fine- and slapping me a fee
---
Could do with some input please.
This only happened after what appeared to be work on the DSLAM the early hours of yesterday morning Thinking.. if they've installed more capacity.. would they have been re-jigging line cards and swapping ports. Have they moved me to a duff port. The line was off for an hour whilst they did what ever they needed to do.
Is there anyway anyone at Plusnet would be able to find out just what they were doing yesterday morning on my DSLAM and why my line was taken out of use for an hour at 04:18.
Cheers.
Despite Interleaving and INP being applied, this has done nothing to arrest the errors and they still continue at the same rate.
I still have the same amount of CRCs and ErrSecs.. I just also now have lots of FECs too... so interleaving has done naff all other than to increase my latency by 7ms.
Ive swapped my router back to the HG612 but that makes no difference either.
I think Im going to have to raise a ticket.. but because the error rate has continued, Im unsure if the DLM will get me again tomorrow.
Im still also loathe to raise it because right now I could do without BT hassle and them saying there's nothing wrong with your line - look it syncs fine- and slapping me a fee
---
Could do with some input please.
This only happened after what appeared to be work on the DSLAM the early hours of yesterday morning Thinking.. if they've installed more capacity.. would they have been re-jigging line cards and swapping ports. Have they moved me to a duff port. The line was off for an hour whilst they did what ever they needed to do.
Is there anyway anyone at Plusnet would be able to find out just what they were doing yesterday morning on my DSLAM and why my line was taken out of use for an hour at 04:18.
Cheers.
The application of Interleaving and INP did absolutely nothing and the errors continued at the same high rate....
So.... the DLM this morning decided to change the interleaving depth and rate limited me to 17Mbps.
.... which has done nothing absolutely nothing, Im still getting SNRM spikes... and still the errors continue to rack up at exactly the same rate as before the DLM first kicked in.
Because Im still seeing such high error rates, I guess theres a strong probability tomorrow it will try again
Cheers Andy... Im not sure what exactly it is they do.. but in the cases of a hot SVLAN, dont they sometimes move a load of customers to a new endpoint in an attempt to distribute the load? My downtime was circa 1 hour.
Im far more familiar with the 20CN/21CN setup rather than FTTC capacity.. but afaik the SVLAN is equivalent to the old VPs on the backhaul?
Is there anyone from Plusnet that can check please if there was PEW at my exchange 17/10/2014 during the wee small hours.
Id be interested to know if this was related to the capacity issues Ive been seeing on the SVLAN.
TIA
So.... the DLM this morning decided to change the interleaving depth and rate limited me to 17Mbps.
.... which has done nothing absolutely nothing, Im still getting SNRM spikes... and still the errors continue to rack up at exactly the same rate as before the DLM first kicked in.
Because Im still seeing such high error rates, I guess theres a strong probability tomorrow it will try again
Quote BTw capacity upgrades won't have anything to do with your connection to the DSLAM.
Cheers Andy... Im not sure what exactly it is they do.. but in the cases of a hot SVLAN, dont they sometimes move a load of customers to a new endpoint in an attempt to distribute the load? My downtime was circa 1 hour.
Im far more familiar with the 20CN/21CN setup rather than FTTC capacity.. but afaik the SVLAN is equivalent to the old VPs on the backhaul?
Is there anyone from Plusnet that can check please if there was PEW at my exchange 17/10/2014 during the wee small hours.
Id be interested to know if this was related to the capacity issues Ive been seeing on the SVLAN.
TIA
Thank you for that link. Now that is interesting, because reading through it reminded me that the week commencing 10th of Sept BT undertook major PEW, whereby they were re-routing their backhaul. I was advised by an ISP rep (not PN) they were in effect remapping different circuits to different RAS.
That link mentions SVLAN moves from one bRAS to another, which is why I suddenly recalled the remapping of RASes in Sept. Im sure you can remember how in days of old that which RAS you were connected via sometimes bore no geographical sense, and this affected latency for quite a lot of EU's.
That work was carried out here on the 11th of Sept and I was quite happy at the time, because overnight my latency decreased from 16ms down to 12ms - which is the lowest Ive ever seen it in all the years Ive had adsl.
With hindsight I started noticing slow speeds on the 17th of Sept.. so you can guess why at first I thought it may have been Plusnet... and subsequently DaveT told me it was my SVLAN running hot... but now Im beginning to wonder if the capacity issues may have occured a few days before I actually noticed.
I guess I'll never know as I didnt noticing really slow speeds until the 17th when PN were had the BGN issue. Im not one to do speed tests very often, so it was only when I physically noticed things taking longer than normal that I did a speedtest.
I had to comment on this though
I knew it was coming, and have done for a few years, but seeing it in print by its proper name of MSiP made me feel quite nostalgic, whereas 'withdrawal of IPStream' didnt have quite the same effect.
Increase in capacity can be a software config change, ie light another segment type of thing. It certainly was with the VP's you never saw any down-time when BT lit say another 10/100MB or whatever. After all an (ethernet) SVLAN is similar to an (ATM) virtual path? If so then increases in capacity and creating new SVLANs on an existing pipe should cause minimal disruption.
I guess the difficulty comes when the whole pipe has reached capacity and they need to install new fibre. In days of old, Gb Fibre backhaul could easily support a DSLAM/MSAN of adsl1/adsl2+ users... and if more capacity was needed then it was highly likely a new MSAN would be required too. So lets just put any new users on the new MSAN.. and if things get really heated and customers complain enough, lets do a lift and shift and put them on another dslam/msan. Win win. The unhappy EU immediately gets put on an MSAN with more capacity - instant result for the EU and it releaves capacity the old MSAN by taking someone off it. Ive been there.
Now this is the bit Im not sure about & how they do it... So what happens with FTTC if a SVLAN is running hot and theres no more capacity in that fibre increase the bandwidth. They presumably are going to have to physically start moving around EUs.. which means down-time. What happens if you have one cab with a pile of heavy users and the link between the node and the dslam is at capacity - supposedly unlikely because not that many users max out at 80Mbps, but its not impossible... and once upon a time they said it was impossible to see congestion on the VPs
Whatever they did.. there is no doubt that last week I was seeing heavy congestion. They did something on Thursday night that cured the congestion. It took them the best part of an hour to complete and its left me with a sick line
That link mentions SVLAN moves from one bRAS to another, which is why I suddenly recalled the remapping of RASes in Sept. Im sure you can remember how in days of old that which RAS you were connected via sometimes bore no geographical sense, and this affected latency for quite a lot of EU's.
That work was carried out here on the 11th of Sept and I was quite happy at the time, because overnight my latency decreased from 16ms down to 12ms - which is the lowest Ive ever seen it in all the years Ive had adsl.
With hindsight I started noticing slow speeds on the 17th of Sept.. so you can guess why at first I thought it may have been Plusnet... and subsequently DaveT told me it was my SVLAN running hot... but now Im beginning to wonder if the capacity issues may have occured a few days before I actually noticed.
I guess I'll never know as I didnt noticing really slow speeds until the 17th when PN were had the BGN issue. Im not one to do speed tests very often, so it was only when I physically noticed things taking longer than normal that I did a speedtest.
I had to comment on this though
Quote MSIP Platform Closure - Planned End Dec 14
I knew it was coming, and have done for a few years, but seeing it in print by its proper name of MSiP made me feel quite nostalgic, whereas 'withdrawal of IPStream' didnt have quite the same effect.
Quote Downtime of 1 hour could be the SVLAN move, although I seem to recall it being quicker than this.
Increase in capacity can be a software config change, ie light another segment type of thing. It certainly was with the VP's you never saw any down-time when BT lit say another 10/100MB or whatever. After all an (ethernet) SVLAN is similar to an (ATM) virtual path? If so then increases in capacity and creating new SVLANs on an existing pipe should cause minimal disruption.
I guess the difficulty comes when the whole pipe has reached capacity and they need to install new fibre. In days of old, Gb Fibre backhaul could easily support a DSLAM/MSAN of adsl1/adsl2+ users... and if more capacity was needed then it was highly likely a new MSAN would be required too. So lets just put any new users on the new MSAN.. and if things get really heated and customers complain enough, lets do a lift and shift and put them on another dslam/msan. Win win. The unhappy EU immediately gets put on an MSAN with more capacity - instant result for the EU and it releaves capacity the old MSAN by taking someone off it. Ive been there.
Now this is the bit Im not sure about & how they do it... So what happens with FTTC if a SVLAN is running hot and theres no more capacity in that fibre increase the bandwidth. They presumably are going to have to physically start moving around EUs.. which means down-time. What happens if you have one cab with a pile of heavy users and the link between the node and the dslam is at capacity - supposedly unlikely because not that many users max out at 80Mbps, but its not impossible... and once upon a time they said it was impossible to see congestion on the VPs
Whatever they did.. there is no doubt that last week I was seeing heavy congestion. They did something on Thursday night that cured the congestion. It took them the best part of an hour to complete and its left me with a sick line
Interestingly, my understanding was BTw/BT TSO were re-developing the 21CN network so that rather than having the 10 BRAS locations across the country, they were moving to the using MSE BRAS (Alcatel 7750) which would be located all over the country (nearer to the exchanges). There was some information here - https://www.btwholesale.com/shared/document/Products/Broadband/17th_April_2013_BB_roadmap_and_MSE_30...
Things are a bit different nowadays!
Out of interest, do you notice that kind of difference when disconnecting/reconnecting to PN? I see a small variation sometimes which has been explained by my traffic taking a different route through the BTw network.
Interestingly, BTw's reports never picked up any issues on that SVLAN. It seemed to go from green to blue at the start of October, but never hit yellow or red.
I think we'll need dave to explain this one.
I know from previous congestion issues, there are situations when new fibres have to be put in but with the advance of technology and fibre, I think hardware upgrades must be another option sometimes. http://community.plus.net/exchange-information/ has mentioned 10G upgrades in the past for some exchanges (technically, it's not the exchange that is congested but the SVLAN).
I wanted to try this for FTTP users in my area - see what happens if we hammer our connections and if we can slow each other down. In theory, we should be able to do this as our bandwidth exceeds the GPON shared capacity.
I have no idea how they move users around, but from previous experience, they've been able to move me onto a different SVLAN and I've not lost PPP sync. So they can do it quickly
Quote from: kitz That link mentions SVLAN moves from one bRAS to another, which is why I suddenly recalled the remapping of RASes in Sept. Im sure you can remember how in days of old that which RAS you were connected via sometimes bore no geographical sense, and this affected latency for quite a lot of EU's.
Things are a bit different nowadays!
Quote from: kitz That work was carried out here on the 11th of Sept and I was quite happy at the time, because overnight my latency decreased from 16ms down to 12ms - which is the lowest Ive ever seen it in all the years Ive had adsl.
Out of interest, do you notice that kind of difference when disconnecting/reconnecting to PN? I see a small variation sometimes which has been explained by my traffic taking a different route through the BTw network.
Quote from: kitz With hindsight I started noticing slow speeds on the 17th of Sept.. so you can guess why at first I thought it may have been Plusnet... and subsequently DaveT told me it was my SVLAN running hot... but now Im beginning to wonder if the capacity issues may have occured a few days before I actually noticed.
I guess I'll never know as I didnt noticing really slow speeds until the 17th when PN were had the BGN issue. Im not one to do speed tests very often, so it was only when I physically noticed things taking longer than normal that I did a speedtest.
Interestingly, BTw's reports never picked up any issues on that SVLAN. It seemed to go from green to blue at the start of October, but never hit yellow or red.
Quote from: kitz Increase in capacity can be a software config change, ie light another segment type of thing. It certainly was with the VP's you never saw any down-time when BT lit say another 10/100MB or whatever. After all an (ethernet) SVLAN is similar to an (ATM) virtual path? If so then increases in capacity and creating new SVLANs on an existing pipe should cause minimal disruption.
Now this is the bit Im not sure about & how they do it... So what happens with FTTC if a SVLAN is running hot and theres no more capacity in that fibre increase the bandwidth. They presumably are going to have to physically start moving around EUs.. which means down-time. What happens if you have one cab with a pile of heavy users and the link between the node and the dslam is at capacity - supposedly unlikely because not that many users max out at 80Mbps, but its not impossible... and once upon a time they said it was impossible to see congestion on the VPs
I think we'll need dave to explain this one.
I know from previous congestion issues, there are situations when new fibres have to be put in but with the advance of technology and fibre, I think hardware upgrades must be another option sometimes. http://community.plus.net/exchange-information/ has mentioned 10G upgrades in the past for some exchanges (technically, it's not the exchange that is congested but the SVLAN).
I wanted to try this for FTTP users in my area - see what happens if we hammer our connections and if we can slow each other down. In theory, we should be able to do this as our bandwidth exceeds the GPON shared capacity.
I have no idea how they move users around, but from previous experience, they've been able to move me onto a different SVLAN and I've not lost PPP sync. So they can do it quickly