[Merged] Manchester Proxies.

[th'engineer]

I am fed up of all this bickering like most customers want it working, not falling over.

It was not meant to be a dig at Bill C or BBKing who i have great respect for, its aimed at certain forum members who jump in on threads critising customers about proxies and other things implying NTL are perfect .

I would have thought the reference to only one particular person in the thread would have explained that.

We all have to act like adults both customers and Engineers and sort out the problems not bicker between ourselves.

[Stuartbe]

Originally Posted by th'engineer

I am fed up of all this bickering like most customers want it working, not falling over.

It was not meant to be a dig at Bill C or BBKing who i have great respect for, its aimed at certain forum members who jump in on threads critising customers about proxies and other things implying NTL are perfect .

I agree - there service is carp and they dont give a hoot about there customers - BUT we must be carefull not to take it out on the NTL forum members... Its not there fault after all - they have a loyalty to the compant they work for and that is a good thing IMHO. It makes them want to help more...

Perhaps its time for the mod team to set up a petition that we could send to the management !!!

[th'engineer]

Great improvement these proxies
Traceroute www.ntlworld.com (62.253.162.30) HopIP AddressHostnameAverage RTT11192.168.1.8baal.pair.net0.37 ms2144.232.248.125sl-gw9-rly-6-0.sprintlink.net5.90 ms3144.232.14.37sl-bb23-rly-3-0.sprintlink.net6.21 ms4144.232.14.133sl-bb21-rly-9-0.sprintlink.net6.22 ms5144.232.20.123sl-bb20-tuk-0-0.sprintlink.net12.45 ms6144.232.20.133sl-bb21-tuk-15-0.sprintlink.net12.57 ms7144.232.19.70sl-bb21-lon-14-0.sprintlink.net81.75 ms8213.206.128.59sl-gw11-lon-15-0.sprintlink.net80.20 ms9213.206.159.246sle-ntl-2-0.sprintlink.net90.72 ms1062.253.185.138win-bb-a-so-300-0.inet.ntl.com94.18 ms1162.253.187.222win-dc-a-v902.inet.ntl.com98.62 ms12Time-out13Time-out14Time-out15Destination host unreachable

[BBKing]

Eng - repeat two hundred times - I Must Not Confuse ICMP Echo Packets With TCP Packets On Port 80.

Then repeat two thousand times - Not All Hosts On The Internet Implement RFC 1122 Correctly.

RFC 1122, page 42:

Every host MUST implement an ICMP Echo server function that receives Echo Requests and sends corresponding Echo Replies.

A host SHOULD also implement an application-layer interface for sending an Echo Request and receiving an Echo Reply, for diagnostic purposes.

Originally Posted by th'engineer

I am fed up of all this bickering like most customers want it working, not falling over.

snip

We all have to act like adults both customers and Engineers and sort out the problems not bicker between ourselves.

Well i picked a right week to give up smoking so yes i am a bit touchy at the moment, But you started the bickering by having a go at someone that is trying to help 10 out of 10 on the good moves table if you ask me.

lets give the poeple that control the systems the chance to check them and fix them if it needs that but having a go at those poeple will not help.

Just my feelings on the subject

[mr-b]

Originally Posted by BBKing

Then repeat two thousand times - Not All Hosts On The Internet Implement RFC 1122 Correctly.

It's not so much that, it's that recent virus outbreaks e.g. the W32.Welchia.Worm, have used ICMP to check for potential hosts to infect, and because of the increased traffic many web hosters and service providers have either turned off ICMP or else drop the packets at any hint of any Denial of Service attacks.
Therefore traceroute and ping are just not reliable indicators of network connectivity.
So there's no substitute for looking at network traces to see exactly what's going on.
Some free tools are www.ethereal.com or its relation http://www.networkchemistry.com/products/packetyzer/

[Chrysalis]

any comments made by me are aimed at ntl as a company and I appreciate ntl staff been here helping out users.

thought I would just say that.

Originally Posted by mr-b

It's not so much that, it's that recent virus outbreaks e.g. the W32.Welchia.Worm, have used ICMP to check for potential hosts to infect, and because of the increased traffic many web hosters and service providers have either turned off ICMP or else drop the packets at any hint of any Denial of Service attacks.
Therefore traceroute and ping are just not reliable indicators of network connectivity.
So there's no substitute for looking at network traces to see exactly what's going on.
Some free tools are www.ethereal.com or its relation http://www.networkchemistry.com/products/packetyzer/

All my servers block ICMP unless it comes from specific IP's (which obviously belong to support staff). So we can ping them and trace to them, but joe public cannot. This is not uncommon these days.

Most people tend to confuse pings and tracerts with web traffic (and tcp/udp traffic in general) - I would love a pound for everyone who posts a tracert when they are having http/https problems. It is an easy mistake to make (the same problem happens with gamers, who post ICMP pings when their games use UDP or TCP).

[th'engineer]

Its allright Me and Bill are still speaking.

[Chrysalis]

hmm I had problems downloading a small file from a site, the file changes every few hours and it turned out because this proxy wasnt keeping it up to date, will ring up ntl again later as they have yet to get back to me on either of my complaints and my billing date is tommorow.

[Chrysalis]

the site in question is to do with firefox, it is a development site for firefox extensions and new extensions get uploaded daily sometimes every few hours and this proxy is causing me all sorts of problems at the moment

I am wondering if its possible to setup a a straight pass thru proxy on a linux box that doesnt actually cache page, but allowing me to bypass ntl's proxies, you think this is possible?


come across this

http://www.squid-cache.org/

only problem is I will be setting it up on a us box, wether this makes things faster and better for me than ntl's local server's I dont know.

[mr-b]

Originally Posted by Chrysalis

the site in question is to do with firefox, it is a development site for firefox extensions and new extensions get uploaded daily sometimes every few hours and this proxy is causing me all sorts of problems at the moment

I am wondering if its possible to setup a a straight pass thru proxy on a linux box that doesnt actually cache page, but allowing me to bypass ntl's proxies, you think this is possible?

I think this has been covered elsewhere in this forum, but the short answer is no, you cannot bypass the upstream proxies by using your own. This is because your proxy will use port 80 (like any browser) to talk to web sites and so it will get redirected as well.

Regarding your problematic site, is it possible to do Ctrl-Refresh (for IE) as shown in Robin Walker's cablemodem help pages?
http://homepage.ntlworld.com/robin.d...l#refreshstale

Also you could enter your problematic web page into the Cacheability Engine to see if it's actually cacheable.
http://www.web-caching.com/cacheability.html

The HTTP headers that are usually used to gauge content freshness are:
Expires, Cache-Control, Last-Modified

If the headers indicate that the content is old, then it won't get its freshness verified as often as new content.

[Chrysalis]

I am not planning on using port 80 for my proxy I am aware port 80 wont bypass but I got the impression that using a different port does bypass ntl.

[mr-b]

Originally Posted by Chrysalis

I am not planning on using port 80 for my proxy I am aware port 80 wont bypass but I got the impression that using a different port does bypass ntl.

There are two different ports - the http proxy port (which is commonly set to 3128 or 8080) which is used for inbound traffic (from your browser) and the outbound http port (which is used to fetch content from web servers on port 80).
Changing the inbound proxy port won't make any difference to the fact that outbound http traffic on port 80 will get redirected.
You can configure your proxy to use a parent proxy on say port 8080 if you want to bypass the redirection process if that is broken, but you will still be using an upstream proxy.

The only real way to completely bypass proxies is to encapsulate your web traffic via an application web tunneling gateway on the internet somewhere. But that in itself can caus all sorts of performance problems with packet fragmentation etc. and you need a server on the internet.

http://www.htthost.com/ may be one.

[Foo Fighter]

Couldn't get to microsoft.com earlier, guessing its ntls proxies rather then ms falling over

Turn them off they are crap!