dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
10548
share rss forum feed


swintec
Premium,VIP
join:2003-12-19
Alfred, ME
kudos:5
reply to kilrathi

Re: [TWC] packet loss / service degraded - amp in rockaway park,

Do you have any trace routes at this point that show packet loss, since the other nights maintenance?



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

2 edits
reply to kilrathi

»mcs2.reliablehosting.com/myspeed···ype=voip

Oh and I spoke with t3 and they also logging me and since thursday when noise was fixed upstream snr showing lowest 27 where before it would drop every night to 20.... however error rate on upstream and only upstream is back on today, and at same time the 2 higher modulation channels keep showing either 64QAM or 16QAM where during day its always 64QAM, t3 guy said that most likely channels are hopping on upstream and may cause this but he sees error rate between modem and plant on the upstream.

I actually have 2 lines ran 1 from each pole/tap on street and tried both same problem right now.

»www.pingtest.net/result/66813391.png

It may not sound bad when its 2-3% packet loss but for gaming wise its horrible especially since its upstream only sometimes you feel like you not even controlling anything further I normally broadcast and cast games for tourneys and other vs games but with this issue the broadcasts keeps cutting off and/or dropping frames.

I may have to give up broadcasting and shoutcasting games which is a shame because it was fun and I actually received some hardware from some tournament sponsors in return.

I may have to also go to lower tw package which is rr turbo 20/2 which uses docsis 2 modem and does not use upstream channel bonding to get rid of the packet loss so I can at least use voice communication and play games without constant problems.

I will call customer rep again tomorrow and maybe the plant tech to see if anything else can be done.

Only thing I have to find out whether rr turbo uses the 2.560 Msym/sec 16qam or 5.120 Msym/sec 64qam channel for upstream, I hope it is the lower modulation one as based on my observations for past 2 months and ever since upstream channel bonding kicked in those higher modulation ones at least here have packet loss issues. Few times I was locked only on the lower modulation channel or when all other T4 time out and when I was locked at only the 16QAM channel for a night there was 0 packet loss problems.

Yes speeds would be big let down from what I am getting now but at least I could game and voice comm without constant problems and packet loss, I could still cast games after I record them and upload them after the fact.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

1 edit
reply to kilrathi

sigh I guess there must be something else causing this packet loss maybe its simply occuring on the node itself since 10pm last night constant packet loss on the upstreadm, now its 5am and its still going:



morning 9 am and now I actually noticed when I traceroute on 5th hop of tw it actually stalls (and no i have name resolution off) and currently packet loss is actually getting worse:



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

3 edits
reply to kilrathi

Not sure whats going on but having hard time browsing and some voip tests or packet loss tests simply freeze.

In addition since last evening the modem is having constant spam of T3 errors (No Ranging Response received - T3 time-out).....

Now friday and saturday night and day everything was fine(after noise was fixed and upstream snr corrected) but this thing that started sunday evening could be maybe something else that just occured maybe on tw network here regionally or node.... not sure whats going on but I hope some time at least a week can go without problems...

Test Type: Voice over IP
Downstream Jitter: 0.5ms
Downstream Loss: 0 %
Upstream Jitter: 0.2ms
Upstream Loss: 5.2 %


It seems this maybe differnet problem than before because its going on into the day now..... before with the noise days were clean but this is going on and on...

Oh and signal levels etc are all same as for past 2 months except upstream snr now never drops below 27 since thursday night fixes(reading provided by t3 techs). So more and more points to something going on in tw network/plant/node or misconfiguration.....



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:16
reply to kilrathi

said by kilrathi:

Only thing I have to find out whether rr turbo uses the 2.560 Msym/sec 16qam or 5.120 Msym/sec 64qam channel for upstream...

Turbo is only a change in the allowed bandwidth in the config file of the modem. All the same upstream channels will be used.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

1 edit

negative, t3 escalation center told me if I downgraded I would have to use docsis2 modem and only 1 channel is used, I asked him last night.

Further, I noticed this:

Red 7/23/2012 2:25:39 PM THE 5700 BLOCK OF SHORE FRONT PKY THE 100 BLOCK OF BEACH 59TH ST THE 5700 BLOCK OF SHORE FRONT PKWY under mainteance page, this is not far from here so I truly hope this is related because today all day I have t3 errors and packet loss, before today I didnt have issues like this during the day.

I spoke with client relations today and old issue is considered closed since back then packet loss was result of upstream snr being out of range, since that was fixed everything was great for 2 days. However this new problem is happening constantly since last night 9pm EST and its been happening entire day, upstream snr is within range(above 27) and all other signal levels are perfect and not changed. Thus bascially this is considered a *new issue* even though its affecting me in similar way a constant upstream packet loss.

Really really dissapointed that I only had 2 days of good service before another problem kicked in. The modem had about 200 or more T3 timeout errors through the day as well....



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:16

said by kilrathi:

negative, t3 escalation center told me if I downgraded I would have to use docsis2 modem and only 1 channel is used, I asked him last night.

True a DOCSIS 2 modem can only use a single upstream and single downstream at a time, but it WILL move between multiple channels as the CMTS commands it. Being DOCSIS 2 it can use either 16QAM or 64QAM channels. So moving to Turbo doesn't avoid your problem, it'll just randomly pop up as the CMTS commands your modem to use that particular degraded upstream channel, just like Wideband and Extreme tiers with DOCSIS 3 before upstream bonding was activated.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

you maybe right is just an idea I have because frankly for 2 months now all I had is problems and now I had 2 days of good net was happy and now again a problem except now its constantly not just during nighttime.... currently packet loss is constant 2-7% and t3 timeout errors constantly, it took me 2 months to get plant tech to fix previous problem with upstream snr but this looks like new problem and i pray is something they are aware of because if I have to wait another 2 months to fix this problem then this is gotta be one of the worst isp experiences I could think of.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi

That issue in far rockaway shows closed now but the packet loss continues, entire day no change.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi

Yesterday jul 23 at 8pm est t3 time out errors stopped and had no issues with packet loss at all, so far entire day today been great. I wasnt aware but plant techs came tonight to let me know they were dispatched to look at my issue, however I had to tell them since yesterday everything is great after whatever was done in far rockaway.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi

I am on the ropes, 1 day everything is great and then next day today constant packet loss again on the upstream....

I spoke with plant tech again and they will try to dispatch someone again with the next shift, only thing I can think of right now is that when one specific channel is used with upstream it has issue while another when its used it doesnt and that could explain the packet loss. Signal levels are all perfect including upstream snr, this is simply something else thats going on maybe even at the node itself.

Constant 4-10% packet loss right now.

There were bunch of T3 timeout errors occuring at around 8pm and seems that time packet loss galore..... and they wonder why I dont call to report the issue is resolved.... well because I have to actually wait few days to be sure and bam same thing different day.



I am actually having problem to browse the internet some pages just stall/timeout.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi


2 lower modulation channels no packet loss
Again, I have a cheap way to verify that the packet loss occurs on one of the higher modulation upstream channels, when I add extra splitter 2 of the higher modulation channels on upstream T4 timeout, thats when I remove the splitter, now the internet connection stays alive because upstream is thrown on one of the other 2 lower modulation upstream channels as shown in the picture.

When those 30.6 and 23.8 Mhz channels are T4 and not used and only the lower modulation channels are used I have no packet loss at all, no problems with the connection and I still hit my downstream/upstream max speeds of my package (50/5).

Of course this may only last half a day or day before modem will reboot but until tw can fix the issue with packet loss on the higher modulation channel this is only way I can try to attain a connection without 5-10% upstream packet loss.


kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi

I spoke with client relations again and plant tech supposed to investigate again but I have low expectations at this point, I assume its very difficult issue where some ingress/egress such as microreflections is occuring thats affecting the higher modulation/symbol rate upstream channels thats causing packet loss. Maybe my upstream tx is too close to noise level (about 37-8db without splitter) and this basically makes those modulations not possible to work without excessive error rate. At this point I am trying to get them to push custom config for me so that I only use the 1/2 lower modulation channels if thats at all possible because clearly with those lower modulation channels there is no issue. There is no need for 50/5 package to use the higher modulation upstrem channels at all and its only causing problem for me. Its basically a sign of tw pushing tech thats not needed for current packages when the system/wiring here is not ready for it. Modem lost dhcp renew ability and had to reboot 12 hours after I did the 2 channel t4 trick. But at this point I will have to keep doing this if I want to have packet loss connection. Sad but this is reality.



Smith6612
Premium,MVM
join:2008-02-01
North Tonawanda, NY
kudos:24

Be nice if they're able to get the higher modulation to work as it makes for more upstream capacity, but you gotta do what you gotta do, I'd say until Timer Warner is able to get the plant cleaned up and fixed a bit more.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

said by Smith6612:

Be nice if they're able to get the higher modulation to work as it makes for more upstream capacity, but you gotta do what you gotta do, I'd say until Timer Warner is able to get the plant cleaned up and fixed a bit more.

Yes, although at this point it seems I may have to reboot and do the "trick" every 12 hours due to ip lease which is very annoying, I will try to pursue plant techs whether maybe there is possibility of custom file for my mac address where it forces to only use 1 of the 2 lower modulation upstream channels, otherwise I will ahve to see if rr turbo docsis 2 uses only the lower modulation channels for upstream as that would be only option.

JCHII75088

join:2012-03-06
Rowlett, TX

1 edit
reply to kilrathi

Post erased.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

2 edits
reply to kilrathi

Correct, however the issue thats causing problem is they using higher modulation upstream channels when there is simply no need for it for even the 5mbit upstream pacakge, and thus its just really bad decision on their part. Basically my current solution is to reboot the modem after 12 hours when ip lease expires and then hook up extra splitters once modem boots and then switch one to 8db so that it causes 2 higher modulation channels to t4 and then quickly switch back so the 2 lower modulation channels stay working, this way modem will not try to properly renew its ip but for 12 hours i can use it without packet loss at all, i been doing this for past few days since this is only way i can have clean connection. I told client rep and plant tech the situation which clearly indicates the packet loss occurs on the higher modulation channels, maybe just 1 of them or both that I am not sure of. What I am sure of is when only those 2 lower modulation channels are used for upstream i have no issues at all.

when I do this the default channel which is one of the 64qam that i force to t4 I believe is responsible for communicating with cmts for stuff like renewing lease etc... so basically I have no way to keep this connection past 12 hours, thats the only downside.

Every time I reboot though I always check and there is always some upstream packet loss when higher modulation channels are locked(all 4). When only 16qam channels are used there is no packet loss on upstream. Its quite clear.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi

I been using my method of T4 timing out atdma channels upstream and keeping only the 16qam tdma channels working and even got it to work without having to reboot every 12 hours, this way I been having 0 problems no packet loss at all and max speeds just fine for past 2 days. I will continue to do this until I cannot or they let me know somethin was done about ingress on higher modulation upstream channels.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi

So something happened tonight where I can no longer block out the 2 higher modulation 64QAM upstream channels and only use the 2 16QAM with my methods, I thought maybe by now they fixed the problems but nope still packet loss on the upstream and its constant at regular intervals when doing voip tests. TW simply doesnt seem to care about some kind of ingress on the line here that simply makes 64qam upstream have packet loss here, I will be attempting to keep rebooting modem adding splitters back and forth and try to have only the 16qam upstream channels working, for 9 days I had it that way and had no issues at all, but once I am forced to use 64qam upstream channels thats when packet loss strikes. I spoke to plant techs 2 weeks ago I was told they will call me with regards if its possible to send custom config for me to only use those but never got call back and never got call back about the issue either. I been quiet since I had it working my way for over a week but it seems now I am unable to do this anymore(for past 2 hours).


nephipower

join:2012-02-20
San Antonio, TX

What a bummer, i'm sorry to hear that man



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi

I got it working again without packet loss, Took about 10 modem reboots and now I only have the 2 16QAM upstrream locked and other 2 are T4. This way there is no packet loss but I have to usually reboot modem every day. I been trying to contact client relations today but assigned person to me was not available supposed to call back. Now if I knew that one of the lower packages uses only those 2 upstream lower modulation channels I would jump on that rigth away but noone knows anything for certain even the plant tech said he wasnt sure.

Wonder if anyone here(NYC/Queens) is on the older roadrunner and turbo packages if you can see your modem stat pages what kind of upstream channel specs do you have?



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:16

said by kilrathi:

Now if I knew that one of the lower packages uses only those 2 upstream lower modulation channels I would jump on that rigth away but noone knows anything for certain even the plant tech said he wasnt sure.

ANY service tier will use any of those 4 available channels, the tier doesn't dictate what channels are used. The modem and CMTS dictate what channels are used. The CMTS will put the modem on whatever upstream channel it needs to be on, that the modem can handle.

The only way to lock out using QAM64 channels on your end (other than putting the upstream channels out of spec for QAM64 like you've been doing) is to use a DOCSIS 1 modem.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

Ahh that is sad news, but do you have any idea if plant tech or engineer would have any capibility to send a custom config file to my modem so that it only used 16qam channels for upstream and others only for qpsk initaliziation etc... Thats what I am trying to find out but its been long process. Sure they could get around and sweep the area and those frequencies and find where the problem is occuring but I dont count on that anytime soon.

I did notice something like this today on network status page:

Red 8/8/2012 7:20:00 PM HUB-E C401 4/10 30.6 MHZ

Red 8/8/2012 6:10:00 PM HUB-F C401 2/9 23.8 MHZ

under Degraded Service Outages
Today!

Now I dont know whether that is related to my problem but both of those frequencies are the 64QAM frequencies used on upstream here where I live and both of them have the constant errors so I hope maybe they finally getting around to do something about this.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY
reply to kilrathi

I spoke with t3 tech today and he said that currently its that 30.6 MHZ channel thats constantly jumping in upstream snr between 20 and 30(this is when I have modem locked normally with all upstream channels not blocked). And since thats the channel id thats default channel so its being used for upstream and yes I have constant packet loss. This upstream snr jumping occuring on that frequency is most likely sign of some noise or interference thats causing the packet loss. I am trying again to get plant tech out there since it seems thats only the frequency thats a problem.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

1 edit
reply to kilrathi

So I was migrated to a new CMTS that has 6 downstream channels and all 4 upstream channels are max 16QAM last wednsday, for 5 days everything was great and I was about to call today to let them know it seems they finally after few months solved my issue when just as I double checked I noticed my connection die as I was opening some sites and so I go check my statistics and boom I see insane packet loss back:



I checked visualtrace voip tests and all that packet loss is occuring on the upstream yet again.... sigh just when I think the nightmare is finally over it strikes again, and actually right now its a lot worse than before.

The worst part of this is that most likely the noise occurs on the same frequenices on upstream channels as before but while before I was able to add splitters to have 2 of those channels T4 and then have the modem still operate on the remaining 2 for at least 24 hours without packet loss now with this new CMTS all upstream channels have QPSK modulation and thus I am unable to "drop out" certain channels anymore for an easy fix and thus if this continues I am back to square one back to same problem that started back at end of may...



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

1 edit
reply to kilrathi

Nothing is still done, client relations rep promised me to call back last week and never did regarding update from engineering. Meanwhile anytime I see "No Ranging Response received - T3 time-out" in the log of the modem thats when packet loss jumps up on the upstream in rates of 5-20% and sometimes connection will freeze out for few minutes but the modem itself will never lose lock, I suspect it just jumps upsteam channel after that to one that has no noise but then goes back, I have pretty much narrow down which of the 4 channels have noise(packet loss) and which dont but with new cmts there is really no way to force specific one to be used as all have QPSK.

Yesterday between 3pm and 1am modem was full of T3 timeout errors and there was upstream packet loss occuring along with degraded upstream speed. Signal stats are perfect but there is noise on 1 or 2 of the upstream channels that causes the problem. There is no way to force a specific upstream channel and thus unless TW gets off their butt and manages to find out cause of the noise on the given channel I am stuck with this problem.

Since memorial day this same problem been going on and nothing is still really done about, if anything migration to new cmts only negated a way I found to solve the problem on temporary basis by myself.

Now this new cmts I been migrated to seems to be in NJ(it has NJ hostname) and I noticed that now the packet loss upstream problems kick in for past few days from about 5pm till 1-2 am so it matches more of peak time usage than packet loss problems on old cmts when it would happen all night and it was noticable even during day, now around 5pm T3 time out errors will start and the upstream is very jumpy it will be fine for 20 seconds meaning full speed and no packet loss and then bam could totally freeze for like 10 seconds or get about 20 % packet loss for a while, this all matches T3 timeout errors on the log modem, when they stop the problems stop too.

Further, there is extra cable line ran to my house from a differnt pole/tap that was left over from when they were trying to see if that will help, now I hooked up my ground block to that one and besides havign different dbmv stats there is no difference in packet loss and thus its definitely not the bad tap that I am hooked to or the line thats running in, further again coaxal from ground block to my modem is brand new no splitters etc...

I am wondering if that new cmts now is serving more customers than when I was migrated to it week ago and now its overloaded, its hard for me to say but so far past few months things are getting basically worse and even level 3 tech support and client relations are not as helpful as they were back in may. Maybe they saw that I am back paying my bill fully and think hey looks like he will just endure the problem, well last bill was last I paid now and unless I get some refund again this will be it. Still cannot get verizon fios unless its the multi dwelling vdsl so I am stuck, but this week I will be calling l3 tech every day and same with client rep I will be asking to get tech out just to drive as its a plant issue or plant tech is needed and reg tech cant do anything but hey if it takes bothering them to have to send tech all the time then I will do what I have to do over and over.


harald

join:2010-10-22
Columbus, OH
kudos:2

The new CMTS if in the same location as the old CMTS, not NJ.

What is your source for knowing that you are on a new CMTS?



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

First hop beyond my modem would be CMTS no? With old one I couldnt even resolve or ping first hop but since CMTS migration there have been 2 different 1st hops, originally a one with hostname that had nj.res in it but after I saw some scheduled work on the queens few days ago on network status page now its a ny.res first hop its kinda weird, but I assume the first hop is cmts am I correct? or at least should be near it.

Either way they had work on sunday and monday on network status page showing HUB-G C### #/# TO #/# ALL UPS FREQS under degraded service and ever since that was removed I have no problems once again for a day now, I noticed they have more ALL UPS FREQS but different hub today so maybe they been doing and are doing major work either way I am glad there are no problems again for a day but as I stated many times before there been times where everything was fine for a day or 2 and then bam, so unless I see this problem gone for like a week or 2 I wont be celebrating.


harald

join:2010-10-22
Columbus, OH
kudos:2

"First hop beyond my modem would be CMTS no?"

No. Modems are layer 2 devices (bridges that route packets based upon hardware addresses) and do not show up in traceroutes.

The address that you've attributed to the CMTS is actually the next router after the CMTS. Unfortunately one cannot rely too much on the location information implied by the naming of routers internal to cable systems. Keeping them up-to-date and accurate is not a priority for a lot of cable companies.

Probably has nothing to do with your problems. Sounds like they may be having some noise problems on the upstream freequencies.



kilrathi
Premium
join:2005-04-22
Rockaway Park, NY

Well yes, I know there is noise on upstream frequencies I know which ones too, the one positive is that new cmts uses max 16QAM on upstream and thus there is less problems in theory with noise as it can operate at lower SNR, with old cmts 2 of the channels were 64QAM and their snr would jump between 18 and 30, with old cmts i could add extra splitters that caused the 2 higher modulation channels to T4 time out due to power level being too high and then quickly remove splitter which caused only the 2 16 QAM channels/frequencies to operate for upstream, this way there was never any problems however I had to reboot the modem usually after a day since I could not renew the ip lease this way.

Based on many calls and the above results both 30.6 and 23.8Mhz frequences is where the noise occured, however T3 did mention the logs showed biggest snr jumps and noise on the 30.6Mhz frequency channel.

Currently with new cmts there been more days with no problems than with problems so I have hope. While I maybe optimistic please remember that the problem was horrible between end of may and until 2 weeks ago when cmts was migrated, and only way I had good connection was with my splitter trick. Aside from 1 month for which I got refund I still paid full 100 bucks for that service so I am more than nice to time warner. I hope that this current pattern of last 2 days holds for long as there was no T3 timeout errors or problems for nearly 2 days now.