PDA

View Full Version : Uploads broken - CS woes

Alan Arndale
31-01-2009, 10:18
Hi all - I'm new here, so I hope you don't mind me weighing in with a tale of woe.

I'm on the 10meg cable service, in HP22 - near Aylesbury, Bucks. For the last 10 days, my upstream has been almost completely dead. Uploads to my FTP server crawl along at about 3-4kB/sec, and eventually time out. If I do a test at speedtest.net, the upload sits there blank for about a minute before registering about 2-3kBsec.

Last weekend, I spoke to Virgin's UK TS guys, and after running me through the usual tests (connecting directly to the modem rather than via a router, trying the connection on other computers, swapping ethernet cables), decided that the cable modem was at fault. A couple of days later I received the cable modem, activated it, and the upstream was still useless.

I called TS again, this time in India. The guy tried everything he could in the script to get me off the phone quick - "Oh you're using a mac" (no, it happens on the two windows PCs here as well), "You're running security software" (no, I'm not), "you need to leave the router unplugged from the modem for a week before you get a true speed reading" (well that sounds like crazy talk, but regardless - it still did it if I connected directly to the new modem). After I insisted that this was a physical problem, he put me on hold. Then he came back and said there's a problem in my area - which is strange, as when I spoke to the UK tech guys, they said this area was running normally. When I pressed him further, he appeared to be saying it was a contention problem, which seems odd when it's only the upload that's affected? He claimed that it'll be fixed on the 31st.

So I'm sitting here on the 31st (admittedly it's early), and no change so far. Is there anyone reading this in the same area as me and experiencing similar problems? Does anyone think it's worth me trying to speak to UK TS again? I was wondering if it's a fault on the cable coming into my house - maybe an attenuator is needed?

Downstream and Upstream data from Modem config :

Downstream Lock : Locked
Downstream Channel Id : 3
Downstream Frequency : 330750000 Hz
Downstream Modulation : QAM64
Downstream Symbol Rate : 6952 Ksym/sec
Downstream Interleave Depth : taps12Increment17
Downstream Receive Power Level : -2.7 dBmV
Downstream SNR : 38.5 dB

Upstream Lock : Locked
Upstream Channel ID : 1
Upstream Frequency : 34000000 Hz
Upstream Modulation : QPSK
Upstream Symbol Rate : 2560 Ksym/sec
Upstream transmit Power Level : 44.5 dBmV
Upstream Mini-Slot Size : 2

And here are three traceroutes to bbc.co.uk from the other night (connected directly to the modem) :

traceroute to bbc.co.uk (212.58.224.138), 64 hops max, 40 byte packets
1 10.213.16.1 (10.213.16.1) 7.543 ms 7.418 ms 11.773 ms
2 wtfd-t2cam1-b-v104.network.virginmedia.net (80.3.2.145) 11.913 ms 9.181 ms 7.892 ms
3 watf-t3core-1b-ge-012-0.network.virginmedia.net (195.182.181.21) 19.629 ms 11.158 ms 8.233 ms
4 pop-bb-b-ge-530-0.network.virginmedia.net (212.43.162.142) 10.307 ms 19.593 ms 14.639 ms
5 gfd-bb-a-so-010-0.network.virginmedia.net (213.105.172.137) 15.910 ms 14.913 ms 11.601 ms
6 gfd-bb-b-ae0-0.network.virginmedia.net (213.105.172.6) 13.346 ms 9.886 ms 11.006 ms
7 redb-ic-1-as0-0.network.virginmedia.net (62.253.185.78) 28.497 ms 14.644 ms 12.930 ms
8 212.58.239.249 (212.58.239.249) 27.770 ms 20.903 ms 19.725 ms
9 212.58.238.133 (212.58.238.133) 14.800 ms 15.856 ms 17.431 ms

-traceroute to bbc.co.uk (212.58.224.138), 64 hops max, 40 byte packets
1 10.213.16.1 (10.213.16.1) 10.534 ms 18.389 ms 7.461 ms
2 wtfd-t2cam1-b-v104.network.virginmedia.net (80.3.2.145) 21.641 ms 11.672 ms 12.682 ms
3 watf-t3core-1b-ge-012-0.network.virginmedia.net (195.182.181.21) 11.952 ms 9.202 ms 11.267 ms
4 pop-bb-b-ge-530-0.network.virginmedia.net (212.43.162.142) 12.773 ms 9.336 ms 10.913 ms
5 gfd-bb-a-so-010-0.network.virginmedia.net (213.105.172.137) 12.191 ms 11.525 ms 15.071 ms
6 gfd-bb-b-ae0-0.network.virginmedia.net (213.105.172.6) 14.420 ms 13.692 ms 10.938 ms
7 redb-ic-1-as0-0.network.virginmedia.net (62.253.185.78) 12.331 ms 11.347 ms 19.721 ms
8 212.58.239.249 (212.58.239.249) 12.598 ms 15.007 ms 13.347 ms
9 212.58.238.133 (212.58.238.133) 31.288 ms 201.416 ms 201.640 ms

-traceroute to bbc.co.uk (212.58.224.138), 64 hops max, 40 byte packets
1 10.213.16.1 (10.213.16.1) 22.835 ms 9.399 ms 7.501 ms
2 wtfd-t2cam1-b-v104.network.virginmedia.net (80.3.2.145) 8.038 ms 10.139 ms 19.492 ms
3 watf-t3core-1b-ge-012-0.network.virginmedia.net (195.182.181.21) 15.095 ms 12.964 ms 19.064 ms
4 pop-bb-b-ge-530-0.network.virginmedia.net (212.43.162.142) 13.393 ms 12.662 ms 26.349 ms
5 gfd-bb-a-so-010-0.network.virginmedia.net (213.105.172.137) 14.220 ms 16.893 ms 13.961 ms
6 gfd-bb-b-ae0-0.network.virginmedia.net (213.105.172.6) 18.223 ms 20.634 ms 25.822 ms
7 redb-ic-1-as0-0.network.virginmedia.net (62.253.185.78) 25.956 ms 31.513 ms 36.327 ms
8 212.58.239.249 (212.58.239.249) 26.903 ms 26.465 ms 16.004 ms
9 212.58.238.133 (212.58.238.133) 13.565 ms 13.157 ms 14.216 ms
Joxer
31-01-2009, 12:39
Ex NTL or Telewest? And which modem? Looks like an Ambit...
Alan Arndale
31-01-2009, 12:47
I presume it's ex-NTL, as my ISP email is @ntlworld.com (only been here 18 months, and it was already VM by then).

Modem is Ambit E08C013.00
Joxer
31-01-2009, 13:12
Everything looks OK to me, I would suspect an upstream SNR issue but the traces look fine any issue would show some packet loss or more erratic times. Upstream utilisation issues would show similar symptons.
Alan Arndale
31-01-2009, 13:26
When they claim that the problem will be "fixed" today, what do you think they might be referring to? I mean, if there's someone (or a number of people) hogging the upstream, does this mean they're going to be cut off?

Incidentally, I forgot to post this earlier :

https://www.cableforum.co.uk/images/local/2009/01/1.jpg

:(
Zhadnost
31-01-2009, 20:08
I think the speedtest result is a bit suspect, 18kb should not be able to support 12MBit.
Alan Arndale
31-01-2009, 21:29
I think the speedtest result is a bit suspect, 18kb should not be able to support 12MBit.

Well, uploads to the FTP for my website are similarly slow, and often time out. I can't play any Xbox Live or other online games, although I can sign in. There's definitely a problem with the upstream.

I've tried running the SWITCH NDT test using the server at CERN in Switzerland. Results as follows :

WEB100 Enabled Statistics:
Checking for Middleboxes . . . . . . . . . . . . . . . . . . Done
checking for firewalls . . . . . . . . . . . . . . . . . . . Done
running 10s outbound test (client-to-server [C2S]) . . . . . 33.0kb/s
running 10s inbound test (server-to-client [S2C]) . . . . . . 9.40Mb/s

------ Client System Details ------
OS data: Name = Mac OS X, Architecture = i386, Version = 10.5.5
Java data: Vendor = Apple Inc., Version = 1.5.0_16

------ Web100 Detailed Analysis ------
Interprocess communications failed, unknown link type.
Link set to Full Duplex mode
Information: throughput is limited by other network traffic.
Good network cable(s) found
Normal duplex operation found.

Web100 reports the Round trip time = 127.08 msec; the Packet size = 1448 Bytes; and
There were 50 packets retransmitted, 748 duplicate acks received, and 622 SACK blocks received
The connection was idle 0 seconds (0%) of the time
C2S throughput test: Packet queuing detected: 73.12%
S2C throughput test: Packet queuing detected: 33.83%
This connection is network limited 99.61% of the time.

Web100 reports TCP negotiated the optional Performance Settings to:
RFC 2018 Selective Acknowledgment: ON
RFC 896 Nagle Algorithm: ON
RFC 3168 Explicit Congestion Notification: OFF
RFC 1323 Time Stamping: ON
RFC 1323 Window Scaling: ON

Server 'ndt.switch.ch' is not behind a firewall. [Connection to the ephemeral port was successful]
Client is probably behind a firewall. [Connection to the ephemeral port failed]
Information: Network Middlebox is modifying MSS variable
Server IP addresses are preserved End-to-End
Information: Network Address Translation (NAT) box is modifying the Client's IP address
Server says [MOD EDIT IP REMOVED] but Client says [192.168.2.151]




WEB100 Kernel Variables:
Client: localhost/127.0.0.1
CurMSS: 1448
X_Rcvbuf: 16777216
X_Sndbuf: 8388608
AckPktsIn: 2960
AckPktsOut: 0
BytesRetrans: 72400
CongAvoid: 0
CongestionOverCount: 0
CongestionSignals: 3
CountRTT: 2213
CurCwnd: 170864
CurRTO: 345
CurRwinRcvd: 524280
CurRwinSent: 17920
CurSsthresh: 149144
DSACKDups: 0
DataBytesIn: 0
DataBytesOut: 13107784
DataPktsIn: 0
DataPktsOut: 8989
DupAcksIn: 748
ECNEnabled: 0
FastRetran: 3
MaxCwnd: 262088
MaxMSS: 1448
MaxRTO: 418
MaxRTT: 249
MaxRwinRcvd: 524280
MaxRwinSent: 17920
MaxSsthresh: 208512
MinMSS: 1448
MinRTO: 245
MinRTT: 42
MinRwinRcvd: 522728
MinRwinSent: 17896
NagleEnabled: 1
OtherReductions: 32
PktsIn: 2960
PktsOut: 8989
PktsRetrans: 50
RcvWinScale: 8
SACKEnabled: 3
SACKsRcvd: 622
SendStall: 0
SlowStart: 0
SampleRTT: 147
SmoothedRTT: 142
SndWinScale: 3
SndLimTimeRwin: 0
SndLimTimeCwnd: 11000023
SndLimTimeSender: 43032
SndLimTransRwin: 0
SndLimTransCwnd: 1
SndLimTransSender: 1
SndLimBytesRwin: 0
SndLimBytesCwnd: 13107784
SndLimBytesSender: 0
SubsequentTimeouts: 0
SumRTT: 281218
Timeouts: 0
TimestampsEnabled: 1
WinScaleRcvd: 3
WinScaleSent: 8
DupAcksOut: 0
StartTimeUsec: 879449
Duration: 11044785
c2sData: -1
c2sAck: -1
s2cData: -1
s2cAck: -1
half_duplex: 0
link: 0
congestion: 1
bad_cable: 0
mismatch: 0
spd: 9.50
bw: 4.76
loss: 0.000333741
avgrtt: 127.08
waitsec: 0.00
timesec: 11.00
order: 0.2527
rwintime: 0.0000
sendtime: 0.0039
cwndtime: 0.9961
rwin: 3.9999
swin: 64.0000
cwin: 1.9996
rttsec: 0.127075
Sndbuf: 8388608
aspd: 0.00000
CWND-Limited: 31.29
minCWNDpeak: 50680
maxCWNDpeak: 262088
CWNDpeaks: 11

The theoretical network limit is 4.76 Mbps
The NDT server has a 4096.0 KByte buffer which limits the throughput to 503.63 Mbps
Your PC/Workstation has a 511.0 KByte buffer which limits the throughput to 31.47 Mbps
The network based flow control limits the throughput to 15.73 Mbps

Client Data reports link is 'System Fault', Client Acks report link is 'System Fault'
Server Data reports link is 'System Fault', Server Acks report link is 'System Fault'
Ignitionnet
31-01-2009, 22:01
What router are you using dude? You should have a packet size of 1460 bytes on Virgin.
Alan Arndale
31-01-2009, 22:17
It's an Edimax BR6204WG. Although I should add that the upstream problem happens when I'm plugged straight into the modem.

I'd rather not switch my internet service providings to a provider providing a better level of service providings if possible ;)

Much love, g'night! ;)
Ignitionnet
31-01-2009, 22:20
I think you'd love a better level of service providings!

Speak with the lovely people on the newsgroups see if they can offer any inspiration as far as the quality of your upstream signals goes.
Alan Arndale
31-01-2009, 22:28
I think you'd love a better level of service providings!

Speak with the lovely people on the newsgroups see if they can offer any inspiration as far as the quality of your upstream signals goes.

Which newsgroups are these? Apologies for my ignorance - new round here, and haven't been on Usenet in about 5 years :blush:
Ignitionnet
31-01-2009, 22:38
Server: news.virginmedia.com
Group: virginmedia.support.broadband.cable
Alan Arndale
31-01-2009, 23:11
Thanks for that. Have posted it up. Fingers crossed.

PS - is the packet size something I can set at this end?
Ignitionnet
31-01-2009, 23:20
Yeah - should be at 1460.

Grab http://www.dslreports.com/drtcp and change MTU to 1460. It would appear to be 1448 at the moment.
Kymmy
31-01-2009, 23:31
External IP removed from the [code] section of post #7
Alan Arndale
31-01-2009, 23:33
Sorry Kymmy - my mistake!
cook1984
01-02-2009, 00:23
I'd say it's unlikely to be an issue with the MTU since all modern operating systems have automatic detection of it.

It sounds like you have ruled out all possibilities to do with your PCs and router, but it might just be worth trying a Linux Live distro just to verify.

Unfortunately it sounds like there is a fault on your line or in your area. Getting VM to acknowledge this is impossible - there is no way to report a broadband fault. All you can do is keep trying tech support and asking them to escalate it or to speak to someone higher up until they make an effort to investigate. Be prepared to spend many hours on the phone over at least several weeks.

One tactic that can help speed things up is to demand compensation of lack of service. Clearly your broadband is unusable. The moment they start loosing money they will take notice. If you pay by Direct Debit, cancel or put it on hold until the fault is resolved.

It might also be worth calling retentions and telling them that you want to leave because the Indians can't fix the fault and you are fed up of listening to their scripted responses. Sometimes they can be convinced to put you through to a UK tech.