Virginmedia: Sites not being accessed -- Please List here

[Toto]

Originally Posted by jarf

The reason your connection works fine is because the route it took to get to that address did not lie across the two Virgin machines I identified in my last post as being part of the problem:

7 12 ms 18 ms 12 ms gfd-bb-b-ge-000-0.network.virginmedia.net [213.105.172.6]
8 12 ms 11 ms 15 ms gfd-bb-a-ge-000-0.network.virginmedia.net [213.105.172.5]

Once the packet gets between those it just gets bounced between them infinitely until they both just give up.

Once again, I state that Virgin have a serious problem somewhere in their network, and will likely cause random sites to stop working for certain machines if they happen to be trying to route packets through those addresses.

Yup, your traces would seem to suggest an internal VM network ping pong game.

[cybernetic_tiger]

Nice... core routing loop on the backbone network at Guildford... I will take a look when I get to work tonight.

[NeilH]

Nice to see even large companies like virgin still suffer issues like this. Seen some drunk admins over the years cause problems like this lol. Anyways, for those of you who aint to familar with routing loops or want the basic idea then visit here.

http://en.wikipedia.org/wiki/Routing_loop_problem

Neil

[jarf]

Well, I called the tech support to let them know, sent the trace report. But still seems to be an issue with the routing loop at the moment.

[cybernetic_tiger]

There is no need to call; it's being looked into.

[KrK]

Jarf isn't the only one with problems in Guildford. I'm also in the same area and experiencing issues with accessing Power Usenet. A quick traceroute gives pretty similar results, with a routing loop cropping up between the same two servers.

>tracert powerusenet.com
Tracing route to powerusenet.com [217.160.249.92]
over a maximum of 30 hops:
1 5 ms <1 ms <1 ms 192.168.1.1
2 9 ms 13 ms 7 ms 10.129.56.1
3 6 ms 6 ms 7 ms glfd-t2cam1-b-v106.network.virginmedia.net [80.4.30.165]
4 8 ms 7 ms 7 ms glfd-t3core-1b-ge-114-0.network.virginmedia.net [195.182.181.229]
5 12 ms 8 ms 12 ms gfd-bb-b-ge-220-0.network.virginmedia.net [213.105.175.89]
6 6 ms 7 ms 6 ms gfd-bb-a-ge-000-0.network.virginmedia.net [213.105.172.5]
7 8 ms 9 ms 7 ms gfd-bb-b-ge-000-0.network.virginmedia.net [213.105.172.6]
8 9 ms 7 ms 8 ms gfd-bb-a-ge-000-0.network.virginmedia.net [213.105.172.5]
9 8 ms 7 ms 7 ms gfd-bb-b-ge-000-0.network.virginmedia.net [213.105.172.6]
10 7 ms 8 ms 7 ms gfd-bb-a-ge-000-0.network.virginmedia.net [213.105.172.5]
...
30 12 ms 7 ms 8 ms gfd-bb-a-ge-000-0.network.virginmedia.net [213.105.172.5]
Trace complete.

The same problem happens with accessing *.powerusenet.com, but accessing a subdomain by its IP address directly seems to resolve the issue.

Most annoying Good to hear that it's being looked into already though, was on the verge of having to plunge into broadband support again *shudders*.

- Pete

[cybernetic_tiger]

This should now be sorted, can you retest please and post any issues??

[KrK]

First off, many thanks for helping out cybernetic_tiger, it's greatly appreciated!

I've just tested it again and it seems that the problem hasn't gone away, but has changed slightly. When trying to accessing "powerusenet.com" or www.powerusenet.com, a tracert now results in the following:

>tracert www.powerusenet.com
Tracing route to powerusenet.com [217.160.249.92]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms 192.168.1.1
2 7 ms 6 ms 8 ms 10.129.56.1
3 9 ms 7 ms 7 ms glfd-t2cam1-b-v106.network.virginmedia.net [80.4.30.165]
4 10 ms 15 ms 8 ms glfd-t3core-1b-ge-114-0.network.virginmedia.net [195.182.181.229]
5 7 ms 7 ms 6 ms gfd-bb-b-ge-220-0.network.virginmedia.net [213.105.175.89]
6 21 ms 9 ms 8 ms win-bb-a-so-010-0.network.virginmedia.net [213.105.172.129]
7 33 ms 12 ms 8 ms gfd-bb-b-so-100-0.network.virginmedia.net [213.105.172.130]
8 12 ms 9 ms 8 ms win-bb-a-so-010-0.network.virginmedia.net [213.105.172.129]
9 9 ms 9 ms 9 ms gfd-bb-b-so-100-0.network.virginmedia.net [213.105.172.130]
10 10 ms 11 ms 14 ms win-bb-a-so-010-0.network.virginmedia.net [213.105.172.129]
...
30 17 ms 15 ms 16 ms win-bb-a-so-010-0.network.virginmedia.net [213.105.172.129]

Trace complete.

It now seems to be bouncing back and forth between Guildford and Winchester(?) Interestingly though, it seems to work fine when you try "news.powerusenet.com", as you can see in this tracert:

>tracert news.powerusenet.com
Tracing route to news.powerusenet.com [216.196.97.161]
over a maximum of 30 hops:
1 <1 ms 1 ms <1 ms 192.168.1.1
2 6 ms 15 ms 11 ms 10.129.56.1
3 19 ms 12 ms 12 ms glfd-t2cam1-a-v106.network.virginmedia.net [80.4.30.37]
4 13 ms 9 ms 7 ms glfd-t3core-1a-ge-115-0.network.virginmedia.net[195.182.181.233]
5 30 ms 10 ms 16 ms gfd-bb-a-ge-200-0.network.virginmedia.net [213.105.175.85]
6 12 ms 12 ms 15 ms nth-bb-b-so-230-0.network.virginmedia.net [213.105.175.138]
7 26 ms 21 ms 15 ms tele-ic-1-as0-0.network.virginmedia.net [62.253.184.2]
8 23 ms 27 ms 22 ms tge8-1.fr3.lon.llnw.net [87.248.208.129]
9 82 ms 88 ms 96 ms tge7-2.fr3.lga.llnw.net [69.28.171.125]
10 95 ms 100 ms 86 ms tge2-4.fr3.iad.llnw.net [69.28.171.157]
11 100 ms 85 ms 86 ms giganews.ge3-4.fr3.iad.llnw.net [69.28.156.174]
12 * * * Request timed out.
...
21 * * * Request timed out.
22 92 ms 86 ms 85 ms news.powerusenet.com [216.196.97.161]
Trace complete.

For curiosities sake, I ran a batch of 10 traces to the news and www subdomains of powerusenet.com, both of which produced the same results as above. The only thing that proves though, is at least its not a pesky intermittent fault

The only thing I can notice is that when it starts looping, it happens when it hits "gfd-bb-b-ge-220-0.network.virginmedia.net" rather than "gfd-bb-a-ge-200-0.network.virginmedia.net", where it sems to work fine.

Hopefully the above traces might be of some use If there's anything else I can try for you that might be of use, feel free to shout

Thanks again,

- Pete

[KrK]

I've just tried a traceroute to www.powerusenet.com, and it seems to be routing to the destination correctly now, as do the sites that Jarf mentioned previously. Looks like whatever was causing the fault has now been fixed successfully.

Many thanks cybernetic_tiger! Your help has been greatly appreciated in all this

- Pete