HTTP and RSync problems

isf · 03-04-2006, 23:49

For the past week I've been experiencing intermittent network problems, on http, ssh and rsync. HTTP downloads have slowed to a crawl and web browsing can be worse than dialup. This all started a week ago, has anybody else noticed this?

Stuart · 04-04-2006, 00:24

to the forums.

How are your computers attached to the network? Cable Modem? STB?

isf · 04-04-2006, 00:47

Hi,

Connected via NAT router to the STB. I originally posted in the traffic shaping thread because I suspect the problems are related to that. I've read the thread about the 27th March maintenance window and the DNS thread, the problems I'm having sound identical to these folk. It's definately not a DNS problem and it's not the proxies, I had it down to traffic filtering until NTL denied it.

Does anybody have any ideas?

Horizon · 04-04-2006, 01:16

Quote:

Originally Posted by isf

For the past week I've been experiencing intermittent network problems, on http, ssh and rsync. HTTP downloads have slowed to a crawl and web browsing can be worse than dialup. This all started a week ago, has anybody else noticed this?

Yes, exactly the same as you for about a week too. Browsing most of the time is near impossible with lots of DNS area box messages appearing all the time. RM post area here in Essex.

Edit: just noticed the DNS thread, probably should have posted there...

isf · 04-04-2006, 01:30

Quote:

Originally Posted by Neptune

Yes, exactly the same as you for about a week too. Browsing most of the time is near impossible with lots of DNS area box messages appearing all the time. RM post area here in Essex.

I've seen no DNS problems, but connections just seem to hang during peak hours. Are the DNS messages you mention from IE?

FWIW I'm routed through Baguley, Manchester.

Siuko · 04-04-2006, 14:14

I find it very odd that me and alot of other poeple seem to have being having very similar issues since the maintenance/upgrade or whatever it was that happened last tuesday and thursday.

I am in the derby area and the person above is in the Baguley area. That is 2 places that were mentioned on the status page when the upgrades were to occur.

So my question is this - if poeple in these areas are having the same sort of problems and since the same time is it not obvious to NTL to connect the two occurences together?

Why must I continually phone Tech Support and be told its - my router, my cables, my pc... my (insert anything besides ntl's infrastructure) causing the problems?

I try to stay calm when these things happen but if its very obvious to me as a NONE NTL person then should it not be VERY VERY obvious to ntl people?

Please lets see some positive action.... I'm willing to help if NTL are stuck! I'm sure it wont be that complicated to pin-point the problems!

Gee · 05-04-2006, 01:01

me too im having same problems im in the sunderland area

isf · 06-04-2006, 14:45

For over an hour, I couldn't even load the forum index last night. I did manage to ssh into another box and bring it up in lynx before that connection locked up. Also managed to post to a thread where someone has been experiencing contention issues since NTL's recent upgrade. Over-contention (saturation) may explain what's happening but I'm only seeing moderate ICMP packet loss. Either way, my home connection is useless.

Has anybody tried reporting this to CS? As you can probably tell, I'm trying to avoid having to deal with them.

IanUK · 06-04-2006, 15:19

Quote:

Originally Posted by Neptune

Yes, exactly the same as you for about a week too. Browsing most of the time is near impossible with lots of DNS area box messages appearing all the time. RM post area here in Essex.

Edit: just noticed the DNS thread, probably should have posted there...

I'm in the RM post area as well, just come back from a weeks holiday and the connection is awful, mega slow and today I'm getting disconnections as well, first time ever

Can anyone techy from NTL take a look and see what has happened please ?

Thanks

Ian

isf · 06-04-2006, 20:03

Quote:

isf@one ~ $ time `wget http://www.cableforum.co.uk/board/forumdisplay.php?f=12`
--18:38:14-- http://www.cableforum.co.uk/board/forumdisplay.php?f=12
=> `forumdisplay.php?f=12'
Resolving www.cableforum.co.uk... 212.227.92.125
Connecting to www.cableforum.co.uk|212.227.92.125|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: unspecified [text/html]

[ <=> ] 78,044 2.33K/s

18:44:15 (221.84 B/s) - `forumdisplay.php?f=12' saved [78044]

real 6m0.821s
user 0m0.003s
sys 0m0.002s
isf@one ~ $

6 minutes to get the forum index (just html without images) yet posting this now, the connection appears to be fine. Looks like I'm going to have to make that dreaded CS call this evening, does anybody have any ideas?

IanUK · 06-04-2006, 20:16

Have you tried specifying NTL's DNS servers manually ?
I found that for some reason it fixed the large pauses.

194.168.4.100
194.168.8.100

Might be worth a try ?

isf · 06-04-2006, 20:38

Quote:

Originally Posted by IanUK

Have you tried specifying NTL's DNS servers manually ?
I found that for some reason it fixed the large pauses.

194.168.4.100
194.168.8.100

Might be worth a try ?

Hi,

I'm not having any problems with DNS, the servers you listed are my pri/sec DNS servers and everything resolves quickly. The problem appears to be specific to either TCP or sustained data transfer.

Thanks for the suggestion.

isf · 06-04-2006, 22:50

Lousy packet loss! This was the first thing I tested but my initial tests were all returning 2-4% which isn't exactly stellar to begin with.

Quote:

1 LAN ROUTER 1.217 ms 1.194 ms 1.179 ms
2 STB (x.x.x.x) 13.040 ms 13.760 ms 34.617 ms
3 bagu-t2cam1-a-v119.inet.ntl.com (80.5.162.77) 9.406 ms 37.754 ms 13.239 ms

Quote:

isf@one ~ $ ping -nqc 50 80.5.162.77
PING 80.5.162.77 (80.5.162.77) 56(84) bytes of data.

--- 80.5.162.77 ping statistics ---
50 packets transmitted, 44 received, 12% packet loss, time 49035ms
rtt min/avg/max/mdev = 9.373/21.734/89.226/15.944 ms
isf@one ~ $ ping -nqc 50 80.5.162.77
PING 80.5.162.77 (80.5.162.77) 56(84) bytes of data.

--- 80.5.162.77 ping statistics ---
50 packets transmitted, 49 received, 2% packet loss, time 49041ms
rtt min/avg/max/mdev = 9.429/19.393/99.489/16.835 ms
isf@one ~ $ ping -nqc 50 80.5.162.77
PING 80.5.162.77 (80.5.162.77) 56(84) bytes of data.

--- 80.5.162.77 ping statistics ---
50 packets transmitted, 45 received, 10% packet loss, time 49037ms
rtt min/avg/max/mdev = 8.800/21.596/154.294/25.275 m

How can NTL not be aware of this?

RonJon · 06-04-2006, 23:12

Quote:

Originally Posted by isf

Lousy packet loss! This was the first thing I tested but my initial tests were all returning 2-4% which isn't exactly stellar to begin with.

How can NTL not be aware of this?

That looks awfull, we need to establish whether its downstream or upstream codeword errors, & where they are occurring. ie is it HFC of is it on NTL's core or outside NTL's network.

Can you do a tracert and PM me with your full postcode & I can check the uBR.

isf · 07-04-2006, 14:36

It's apparently not a HFC issue (thanks RonJon). From work, I'm only seeing MAX 2% packet loss to 80.5.162.77, although I'm regularly seeing max rtt in the region of 2.5 SECONDS.

That machine is the first hop after my STB, can anybody tell me what that box is likely to be?