Loss of DNS? in Leeds 4 times in the last 6 days

~Dan~ · 24-11-2004, 23:15

I've been having problems with my broadband. On friday/sunday/tuesday and tonight. It's done it in the evenings (around 8pm to midnight give or take an hour). It's just come back now at 10.57pm.

Basically what happens is I can access emails, but browsing doesn't work.
I put a url (one based in the USA) into neotrace to see what was going on. It gets from Leeds to London in around 6 or 7 hops, all ip's resolve as they should. The node in London doesn't resolve, then after around 25 seconds of waiting it says the response time on that node is around 20ms and then all nodes after that resolve fine until the destination.

For browsing I can't get pages to work. If I really persevere eventually I can get a page up but it's a lot of hassle and not really worth it.

I'm guessing this means ntl are having problems with a DNS server in the London area? Can anyone confirm this?

EDIT: I've just this minute received 2 emails (from a friend, not spam) that I already received before (one yesterday, one 3 days ago)!

**Paul** · 24-11-2004, 23:49

At a guess you mean the TeleIC routers in London - the lack of response from them is normal - they have better things to do than respond to icmp requests.

~Dan~ · 25-11-2004, 00:32

Quote:

Originally Posted by Paul M

...the lack of response from them is normal...

Whataver was happening I can assure you it's not normal for the net connection to be dead - no webpages loading, intermittent emails.

greencreeper · 25-11-2004, 15:37

Might the times coincide with peak usage - UBR not coping? I await criticism from an NTL tech

ProfPete · 25-11-2004, 19:32

Quote:

Originally Posted by ~Dan~

Basically what happens is I can access emails, but browsing doesn't work.
I put a url (one based in the USA) into neotrace to see what was going on. It gets from Leeds to London in around 6 or 7 hops, all ip's resolve as they should. The node in London doesn't resolve, then after around 25 seconds of waiting it says the response time on that node is around 20ms and then all nodes after that resolve fine until the destination.

Can you post a couple of traceroutes please. Prime candidates are stuff like www.mit.edu and www.harvard.edu for stuff on the american side. www.linx.net will do for this side of the pond.

~Dan~ · 26-11-2004, 03:08

It's working ok now but I'll do that if it does it again.

Matth · 26-11-2004, 20:31

DNS timout for pop.ntlworld.com is a frequent occurence for me, as is other times when a site only goes at the second attempt - are there REALLY only two DNS servers for the entire network, or are the addresses regionally mapped?

steveh · 26-11-2004, 20:45

DNS timeouts began happening again for me in Waltham Forest this week...

**Paul** · 26-11-2004, 20:47

What do you mean by a "DNS timeout" ?

~Dan~ · 26-11-2004, 23:29

It started acting up at 5pm, sites sometimes loading sometimes not, or taking an age to load, with red x's all over the place etc...
It was really bad just before 7pm so I did a couple of traceroutes.

Code:

 
Tracing route to www.mit.edu [18.7.22.83]
over a maximum of 30 hops:
1	16 ms	 8 ms	10 ms 10.76.32.1
(it paused 24sec here)
2	16 ms	 8 ms	32 ms leed-t2cam1-a-ge96.inet.ntl.com [80.0.54.17]
3	12 ms	 *	 10 ms leed-t2core-a-ge-wan64.inet.ntl.com [80.0.241.41]
4	12 ms	14 ms	24 ms lee-bb-a-so-120-0.inet.ntl.com [62.253.185.17]
5	21 ms	 *	 26 ms lee-bb-b-ae0-0.inet.ntl.com [62.253.187.186]
6	33 ms	16 ms	17 ms ren-bb-a-so-000-0.inet.ntl.com [62.253.185.161]
7	30 ms	34 ms	26 ms bre-bb-b-so-200-0.inet.ntl.com [62.253.185.166]
8	27 ms	30 ms	55 ms telc-ic-1-so-210-0.inet.ntl.com [62.253.188.121]
9	35 ms	 *	 30 ms (it paused 22sec here) 195.66.226.185
10 103 ms	96 ms	95 ms p6-0.core02.jfk01.atlas.cogentco.com [154.54.1.5]
11	 *	 98 ms	96 ms p15-0.core01.jfk01.atlas.cogentco.com [66.28.4.165]
12	97 ms	 *	 96 ms p12-0.core01.jfk02.atlas.cogentco.com [66.28.4.10]
13 106 ms 108 ms	 *	 p5-0.core01.bos01.atlas.cogentco.com [66.28.4.117]
14 104 ms 102 ms 106 ms g7.ba21.b002250-1.bos01.atlas.cogentco.com [66.250.14.206]
15	95 ms 106 ms	 *	 mit.demarc.cogentco.com [38.112.2.214]
16	96 ms	 *	 98 ms w92-rtr-1-backbone.mit.edu [18.168.0.25]
17 130 ms 107 ms 107 ms www.mit.edu [18.7.22.83]
Trace complete. 
(total time 74sec)
-----------------
Tracing route to nancy.harvard.edu [128.103.60.24]
over a maximum of 30 hops:
1	11 ms	14 ms	 8 ms 10.76.32.1
(it paused 24 sec here)
2	28 ms	16 ms	16 ms leed-t2cam1-b-ge96.inet.ntl.com [80.0.55.17]
3	14 ms	17 ms	 *	 leed-t2core-b-ge-wan64.inet.ntl.com [80.0.241.169]
4	10 ms	10 ms	 9 ms lee-bb-b-so-120-0.inet.ntl.com [62.253.185.21]
5	17 ms	 *	 48 ms ren-bb-a-so-000-0.inet.ntl.com [62.253.185.161]
6	34 ms	 *	 26 ms bre-bb-b-so-200-0.inet.ntl.com [62.253.185.166]
7	31 ms	39 ms	28 ms telc-ic-1-so-210-0.inet.ntl.com [62.253.188.121]
8	31 ms	29 ms	 *	 (it paused 20sec here) 195.66.226.185
9 102 ms	 *	 98 ms p6-0.core02.jfk01.atlas.cogentco.com [154.54.1.5]
10 106 ms 106 ms	 *	 p15-0.core01.jfk01.atlas.cogentco.com [66.28.4.165]
11 104 ms	98 ms 112 ms p12-0.core01.jfk02.atlas.cogentco.com [66.28.4.10]
12 100 ms 100 ms 100 ms p5-0.core01.bos01.atlas.cogentco.com [66.28.4.117]
13 100 ms 106 ms	 *	 g7.ba21.b002250-1.bos01.atlas.cogentco.com [66.250.14.206]
14	 *		*		*	 Request timed out.
15 116 ms 115 ms 119 ms oxgw2-vl-15-core.net.harvard.edu [128.103.15.15]
16 120 ms 139 ms 116 ms nancy.harvard.edu [128.103.60.24]
Trace complete.
(total time 94sec)

~Dan~ · 26-11-2004, 23:40

Quote:

Originally Posted by Paul M

What do you mean by a "DNS timeout" ?

Whenever you type a url in your browser, for example www.google.com this has to be looked up on a dns server (basically it's like a big address book). The dns server tells you the ip of the url (for google it's 216.239.51.99). Without knowing the ip you (or your pc) wouldn't know whereabouts in the world the site was hosted on and so wouldn't be able to find it. DNS timeouts are where the dns server doesn't (for whatever reason) tell you the ip, so you can't go to the site.
Erm, I think this explanation is right, someone correct me if it's not.

I *think* that if you have the ip's of sites in your hosts file then you don't need a dns server, but obviously without using the dns server you can only visit the sites who's ip's you've already put into the hosts file.

Viras · 27-11-2004, 00:15

Hey all new user here.
Well Dave i can say i've been having the exact same problem as you for the past 4days now, at 5pm im not able to view any webpages what so ever. Infact im not allowed todo anything that involves the internet coz it just doesnt work.

The same pattern over and over again now for the past 4 days have been.

5pm i start getting the problem, then at 10:30/11:00pm my connection becomes stable again and everything works fine.

At first i thought "ok must be planned maintenance" so i thought nothing of it but today [Friday] it happened again so i decided to give ntl a ring only to be on hold for 40mins listening to that damn horrible music

. I then gaveup coz i just couldn't be arsed waiting any longer :P

So i went to watch some tv for awhile to come back at 11pm to find everythings working fine, so i decided to do a little search to see if anyone else is actually getting the same problem as me and i found this forum

I seriously dont know wtf is going on here and ill prolly try giving ntl another ring tomorrow and pray that im not waiting for another 30-40mins

PS, i do live in Leeds and have been having the problem since monday

**Paul** · 27-11-2004, 00:17

Quote:

Originally Posted by ~Dan~

Whenever you type a url in your browser, for example www.google.com this has to be looked up on a dns server (basically it's like a big address book). The dns server tells you the ip of the url (for google it's 216.239.51.99). Without knowing the ip you (or your pc) wouldn't know whereabouts in the world the site was hosted on and so wouldn't be able to find it. DNS timeouts are where the dns server doesn't (for whatever reason) tell you the ip, so you can't go to the site.
Erm, I think this explanation is right, someone correct me if it's not.

LOL

Sorry, perhaps I should have been clearer and saved you a (quite good) explanation. I know what dns is (rather well in fact). I was trying to find out what problem they were experiencing, that they believed was a dns timeout. It is very rare for a dns request to timeout, and since most people don't understand dns they attribute other faults to "dns timeouts". A bit like many people blame the "proxy" server for problems that arn't actually proxy related.

Chris W · 27-11-2004, 00:55

@ Dan-

the traceroutes that you have posted show nothing about dns problems- every hop except for one resolved.

the one where you say it hung for 22seconds (195.66.226.185) doesn't have a domain name associated with it, so it is usual for there to be a pause while this is discovered.

The traceroutes seem to indicate packet loss problems, not dns problems- and pages loading slowly/ with bits missing would tie in with this.

A traceroute is not a good way of testing packet loss because of the way it runs. try pinging a host over a long period instead- try this at a command prompt:

ping 212.58.224.121 -n 100

and then post back the summary statistics. If you see any packet loss, try:

pathping 212.58.224.121 -n

and post back the results from that too.

@ Viras- PM me the mac address of your modem and i'll have a look at what is going on with your connection tomorrow.

~Dan~ · 27-11-2004, 01:36

Quote:

Originally Posted by monkeybreath

@ Dan-

the traceroutes that you have posted show nothing about dns problems- every hop except for one resolved.

the one where you say it hung for 22seconds (195.66.226.185) doesn't have a domain name associated with it, so it is usual for there to be a pause while this is discovered.

Maybe I shouldn't have mentioned DNS in the topic. I wasn't sure what the problem was and that was just me grasping at straws.

I'll try the packet loss thing next time it goes wrong. So that's probably tomorrow then.