Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[kitz]

I know.

Interleaving has increased my latency by about 7ms..  but done absolutley nothing to arrest the CRCs...  which is still the same/slightly worse than yesterday.   
I guess we shall see what happens this evening. 

[Chrysalis]

yeah you was under the 2880/day on that graph, but the errors continued on so I think you got over 2880.

You got some stats from your modem on hourly rate?

Also not sure if the hg612 will save you, if there is a fault or interference any modem will get errors.

[Chrysalis]

interesting is that interleaving with the lower snrm seems to be no better than your higher snrm on fast path.  Both seem to have equal effect of error mitigation.

My line is also still having US snrm dips but isnt causing errors.

[kitz]

interesting is that interleaving with the lower snrm seems to be no better than your higher snrm on fast path.  Both seem to have equal effect of error mitigation.

Yep.. I noticed that and said that the application of Interleaving has done nothing to stop the errors.

Put the HG612 back on - that makes no difference either. Still seeing the upstream SNRm spiking and also similar error rates - in fact if anything there is wider variance between the upstream SNRm band with the HG612 and its done a mean spike down to 1dB...  so its not the Zyxel.

[kitz]

Thinking...  if there was a capacity issue.. and you know how I said the other day that if they were just lighting more bandwidth, then normally its just a software config.

However... if it is backhaul capacity..  I dont see any reason why they should need to switch off the DSLAM for an hour..   unless its to hook up newly blown fibre.   In general if new capacity is installed it should be more of a software configuration..  as the SVLAN is the equivalent of an old VP.

So what if it was to hook up new fibre..  wouldnt they have to do some rejigging in the DSLAM ie move some lines about so that the new capacity is shared??   So what if Ive been moved on to another port on a different line card - which is duff

I dunno what they do.

Looks like Im going to have to raise a fault..   which is going to be a PITA..  because I know damn well what BT Openreach will say.   Theres nowt wrong with your line your syncing at 80 Mbps.. whats the problem?   

[Chrysalis]

kitz you can check if its specific tones with the problem? on the billion tones can be turned off.

So buy a 8800nl, turn the bad tones off and get your fast path back maybe?

I would report the fault, I suspect you will get some crappy engineers initially who will walk after seeing 80/20 and the JDSU passes, then plusnet will end up escalating it before someone with the head on right turns up and fixes it.

But I got a feeling you may get a HR fault showing up on the JDSU,

I know engineers can request a DLM reset before they turn up as my last one did it, as he said DLM can mask problems (which he is correct on) but in your case DLM is failing to mask it.

[kitz]

kitz you can check if its specific tones with the problem?

Hard to say where its occurring. U1 and U2 is spiking 5-6 dB.

I would report the fault,

I was going to do so last night, but the PN fault reporting was undergoing maintenance.  I shall do in a mo.

DLM can mask problems (which he is correct on) but in your case DLM is failing to mask it.

The DLM kicked in again this morning.  It adjusted interleaving depth and has rate limited me to 17Mbps.

Yet again this has done nothing to arrest the error rate which continue exactly the same as before.

[Chrysalis]

I lost sync this morning, alerted by tony's system when I checked my email   After the event DS attainable was shuffling about.

[Chrysalis]

The rate limiting is good in my opinion, thats harder for engineers to ignore.  They may still babble on about it been a a "up to" service but your line spec should be an easy 20 up.

[kitz]

The rate limiting is good in my opinion, thats harder for engineers to ignore.

Yes thats true.   Im also waiting to see what tomorrow brings.  Any action applied by the DLM so far hasnt made any difference to the amount of errors Im seeing, so Im not sure if its finished with me yet.

If you think the SNRm graph looks spikey, thats nothing to what I see on the SNRm per band - thats real pretty - not 
I wonder if the SNRm bands is something that could perhaps be something for Tony to think about if he wants to develop things further.


My CRCs arent any better.  Swapping to the HG612 made no difference and in fact Ive put the Zyxel back on, because
1) the line performed slightly better with the Zyxel
2) Two less Ethernet leads dangling across the desk, 1 less power socket in use  and 1 less piece of hardware.

I dont seem to be able to view ErrSecs on mydslstats for 3 days.  It shows my downstream.. but not any of my upstream.   Can you see them.
From looking at DSLstats they dont appear to be any better either.

[les-70]

  It only seems possible to see your upstream errsecs on mydslwebstats for 24 hours and less. For longer times the scales seem adjusted but nothing shows.

[kitz]

Another day.. another DLM.

Interleave depth 325 to 355,
INP  3 to 3.5
Banding 17000 and now capped at 14994

.. and still the errors keep a coming.

What I noticed is that yesterdays banding gave me a SNRm of apprx 9dB and todays banding gives me an SNRm of 12dB

[Chrysalis]

kitz please keep us updated with ES counters as well.

[les-70]

  mydslweb stats show other oddities that I had not noticed.  Your upsrteam CRC per minute counts seem to be 0,24,46 or 70 almost all the time    --I wonder why other numbers don't see to happen.

    I also wonder what upstream error rate will make the DLM happy.  Your latest reduction in upstream speed suggests that the previous 24 hours did not make it happy and your latest 24 hours look like more than 60/hour but less than 120/hour.  Maybe if it now stays less than 60/hour it will not intervene further. 

    The most interesting thing will be to find out the cause and correction, but I guess just correcting the fault is your priority!!

 

[tbailey2]

  It only seems possible to see your upstream errsecs on mydslwebstats for 24 hours and less. For longer times the scales seem adjusted but nothing shows.

There seems to be a bug in the Column Charts where if the errors exceed a certain value and count, the bars just don't display, especially Upstream. For the moment, and I know everyone is used to viewing bars for ES, the charts for now display as Line Charts like all the other ones but at least all the values are now visible for the longer periods - even if they do look like a map of the Pyrenees at times