[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Slow Internet connection for Fedora 4

Ole Robert Hestvik wrote:
>     Fedora 4 has just been installed. Everything is fine, but the 
> network connection is very slow:
> [root localhost log]# ping vg.no
> PING vg.no ( 56(84) bytes of data.
> 64 bytes from vg.no ( icmp_seq=0 ttl=246 time=3526 ms

I understand that you've got other things to look at at the moment.

Once you've got them sorted, if the problem persists, take a look at 
traceroute vg.no
(I assume that vg.no is not supposed to be local).

[james kendrick ~]$ traceroute vg.no
traceroute to vg.no (, 30 hops max, 46 byte packets
 1  * * *
 2  thus10-hg2.kingston.broadband.bt.net (  22.055 ms
20.751 ms 21.991 ms
 3 (  19.904 ms  21.482 ms
(  147.963 ms
 4 (  23.442 ms
(  23.403 ms (  23.539 ms
 5  park-inside-2-g3-0-0-s195.router.demon.net (  23.429
ms  23.471 ms  21.959 ms
 6  anchor-border-1-g4-0-0.router.demon.net (  23.986 ms
23.479 ms  23.994 ms
17 (  260.106 ms  51.370 ms  51.991 ms
18 (  50.066 ms  51.501 ms  49.969 ms

You should do this from your system.

You get a list of (most) routers between you and vg.no, in order. The
first number in each line is which "hop" it is, followed by the name and
IP address of the router, followed by how long it takes a packet to get
there and back. Traceroute tries three packets: an asterisk means that
the packet got lost.

Of course, it's doing this over the live Internet. It does each line one
after the other. This does mean that if there's a sudden transient delay
over hop 5 when traceroute's looking at hop 13, then hop 13's numbers
will be inflated. So it's not foolproof.

But it should help you tell where the delay is. If it were nearly at the
end, then it's just a problem with vg.net. If there's a 3 second (3000
ms) delay on hop 1, then the delay is between you and your ISP.

Some routers don't support this (my ADSL router, hop 1, doesn't).

What sort of Internet connection do you have, anyway? Dial-up? As this
looks to me like a loaded dial-up connection, or a very loaded other
slow connection. If your system was doing a yum update while trying to
send a couple of big e-mails, and you were surfing the web at the same
time, that would probably do it.

Hope this helps,


PS: For what it's worth, traceroute works by sending ICMP packets with a
low "time to live" field. Each time a packet goes through a router, the
TTL field is decreased by one. If the TTL is zero, the packet is
dropped, and a suitable ICMP reply is sent back to source (so the
sending system knows it's not getting through). This is supposed to stop
infinite loops in the Internet's routing tables.

Traceroute deliberately generates packets with a TTL of first one, then
two, and so on. So the reply tells traceroute the IP of the router, and
traceroute can calculate how long it took for the packet to reach that
router and a reply to be sent.

Normally, TTL is set to something like 255. It's rare for Internet paths
to be longer than thirty hops, so normal traffic shouldn't be affected.

E-mail address: james | ... boxing the books up was a mistake: they are
@westexe.demon.co.uk  | welded to the floor through the power of gravity.
                      |     -- Telsa Gwynne's Diary.

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]