Intermittent Connections / Handshake

[#1]

Evening All,

I've had an issue with my Plusnet connection for several weeks now - namely that I get intermittent traffic flows, particularly when initiating a connection / handshake.

Some services are seemingly quite tolerant of it and work reliably, whilst others (notably BBC content) are particularly intolerant and regularly fail to load, be it audio or video. Several work servers (VPN traffic) struggle to connect.

I write this post plugged directly into the back of the Plusnet router. It's a Hub 1, but I've also tried a TP-Link and Technicolour in it's place - all have the same issue. I've tried different DNS, different MTU's... as things stand now it's as plain as it can be - a fresh factory reset Hub 1.

I've been struggling with Plusnet support, who are insistent at looking at DSL sync logs / line condition issues. An Openreach engineer has already visited, and whilst he did find a voltage leak on the line he was able to clear it and even carried out a lift and shift as an extra precaution - however the issue for me remains.

Given the absolute plain vanilla setup I have at the moment and the clean bill of health on the access circuit, and the fact I can recreate the issue easily - I can only see the issue to be on Plusnet's side, in particular with my specific account. I should also add if I tether out via a mobile operator and repeat my tests I have a 100% success rate.

Is there some kind of traffic inspection going on? Can you rebuild my account on your authentication server (assume RADIUS), or just give me a test one to use to rule out? My support ticket doesn't seem to be going anywhere, so am posting here in the hope that someone else may have experienced similar, or the support staff that man the community areas can assist.

Thanks,

Neil

[#2]

Hi Neil,

A BT Openreach's engineer's "clean bill of health" assurance, is no assurance I would place any reliability on.  At best all that can be trusted is that the engineer reported he found no fault at the time of testing.  That does not mean that there was no fault to find then, nor that a new one has manifested since then.

"carried out a lift and shift as an extra precaution"

What service are you using - ADSL or FTTC?  If FTTC is that FTTC + PTSN or SOGEA?

What was lifted and shifted?  The d-side (which would impact both ADSL and VDSL) or the e-side which only impacts ADSL?

What diagnostic tests have you done?

Which fault bots have you run?  See the links below...

Broadband is not going to work well if the "phone" line is not working properly.

Please perform a quiet line test - dial 17070 select option 2 using a phone plugged into the test socket behind the face plate of the master socket (ideally this should be a corded phone). The line should be silent. A noisy phone line (or no dial tone) will have a marked adverse impact on the performance of broadband.

If the line is noisy or there is no dial tone, then a PHONE LINE fault needs to be raised with your phone provider. If this is PlusNet, you can report a fault via a mobile telephone using the button below.

• Phone Faults

How is your line performing?

Please post your router stats (include line speed, attenuation, SNRM, uptime and if available, error counts); for PN routers use the buttons below, for other routers look at the Kitz link.

Also run a BT Wholesale speed test (ideally using an Ethernet wired PC with the WiFi turned off) then run the further diagnostics, download the .PDF report and post the results here. DO NOT REBOOT the router!

Check the Plusnet user portal for the "Current Line Speed" which is relevant to users with static IP addresses.

• TG582n Stats
• Hub 0 / 1 Stats
• Hub 2 Stats
• Kitz list
• BTw Speedtest
• PN Broadband Status

[#3]

Thanks Townman.

Quiet line test is perfect. No noise.

Using FTTC: PSTN & SOGEA. Lift and shift type wasn't defined on Friday when it was completed so can't be sure which...

3. Firmware version:	Software version 4.7.5.1.83.8.289.1.3 Last updated 28/06/23
4. Board version:	Plusnet Hub One
5. DSL uptime:	0 days, 12:11:20
6. Data rate:	19845 / 67943
7. Maximum data rate:	19845 / 68057
8. Noise margin:	6.0 / 5.8
9. Line attenuation:	22.4 / 17.3
10. Signal attenuation:	22.0 / 17.3

Uptime is low because I carried out a fresh factory reset this morning.

Broadband fault bot led to a support ticket which has been open for some time now.

BTw Speedtest:

Summary of Performance Test

Timestamp

Monday 10 July 2023 | 21:28:39

Ping

15 Ms

Download

63.36 Mbps

Upload

18.66 Mbps

Advance Test

Download Test

Download speed achieved during the test was 63.36 Mbps

Upload Test

Upload speed achieved during the test was 18.66 Mbps

Summary Advanced Test

Your speed test has completed and the results are shown above, however during the test an error occurred while trying to retrieve additional details regarding your service. As a result we are unable to determine if the speed you received during the test is acceptable for your service. Please re-run the test if you require this additional information.

Can't download the report to upload here given the error being seen above.

[#4]

I can see our faults guys have sent you a new router. Have you received that yet?

[#5]

Not yet Will. I'll get it installed as soon as it arrives and report back here and to the support team. To be honest I don't have high expectations as it'll be the fourth VDSL modem / router I've tried, but happy to rule it out.

[#6]

@baldysmurf - unlikely the router stats look good.

@willcutforth - has a hot VLAN been ruled out?  There was a distant history of BTw's VLAN loading reports being over optimistic - is that totally something resolved?

Does the BT speed test system failure have any bearing on performance matters?

[#7]

@Townman had a look at the last weeks data and it's all showing green from the VLAN side. 

It shouldn't have any bearing on it, I've got 72Mbps with no errors etc. Literally nothing on the testing I can find from our side.

[#8]

@willcutforth 

72mbps betwix home and the exchange is one thing, the back haul capacity is a tad different.  We have seen green back hauls before, only to find that in fact they were murky amber!

72mbps is also at variance with the rate reported in the router posted earlier.  has the link speed been variable and if yes, is there any indications of why?

@baldysmurf 

The connected devices for which you report problematic operation ... are they WiFi connected?  Do they support both 2.4GHz and 5GHz bands ... and do both bands have the same SSID.  This is possible on Hub One, but not on Hub Two.

Hub One also works nicely with the BTHH5a variant of router stats - see the link below.  It might be beneficial to run router stats 24x7 with the PnH1 to see if there is anything untoward happening on the line.

[#9]

Thanks to you both for your assistance.

@willcutforth had notification today the Hub Two has just been posted.

@Townman I've ruled out WiFi by connecting via Ethernet straight into the back of whichever router I've been testing with - so currently the Hub One in situ. Good idea on RouterStats - I'll get it setup.

[#10]

@Townman Not by the looks of our lord KBD. Everything is looking smooth.

@baldysmurf  Glad to hear it. Let me know when you receive it.

[#11]

I've had RouterStats running for several hours today whilst the issues have been apparent many times. See log file attached, nothing to note. Attenuation isn't logged there for some reason, but within the application has been solid throughout.

A WireShark capture I've just run has showed several out of order and duplicate acknowledgements. Something is interfering with the traffic and it doesn't look to be the local access.

To recap on the issues:

- intermittent connectivity to multiple destinations.

- websites - initial attempt to load fails, reload will bring up - but then parts of that website may or may not load - for example comments within a news post. Last night took multiple attempts to get 2FA back from my bank for some online transactions.

- video / audio streams - BBC seems worst impacted. iPlayer streams fail to start several times, keep trying and eventually will start and then be fine for as long as you want to stream for.

Some applications are seemingly more tolerant than others, suspect they're retrying more frequently. Teams calls for example are always good, streaming from ITVX or 4OD - always good. Download a large file and will get line rate speeds.

It's just that initial connection that seems to be the problem. Multiple DNS servers already tried.

The new router hasn't arrived yet - I'll post back once it has.

[#12]

@baldysmurf 

With regards to routerstats, you need to change the sample rate to 10 seconds with 720 plotted points per page, which gives a two hour view. Setting are found under the Configure tab. 

You only need to post noise margin graph rather than the text file. If other graphs are needed they can be asked for.

The downside of the Hub Two you will no longer be able to use routerstats

HTH

[#13]

Thanks @Dan_the_Van.

@willcutforth Hub Two arrived Saturday - I hooked it up in bridge mode with my Ubiquiti Dream Router behind. I previously removed the UDR to solely use the Hub One for benefit of keeping things simple. Issues remain - regular connection timeouts as noted previously. I'll remove the UDR again in the coming days to run solely on the Hub Two to see if any improvement, and if not will replace the Hub Two for the One to run routerstats again.

As noted previously, once a stream is going connection holds fine, which if I had a line issue I'd expect to have interruptions during. It's the initial connections that keep failing.

[#14]

Yeah I think with the fact you had multiple routers out made me feel like it wasn't going to make much of a difference but I just wanted to see before doing anything else. If you can try the Hub Two by itself and then see how we are from there.

[#15]

Used the Hub Two exclusively this afternoon @willcutforth, no difference unfortunately.

Let me know if you need RouterStats from the Hub One again, happy to collect and share. Although note previous comments about this being an initiation issue only, once a session is established it's stable.