|
Advanced ADSL Troubleshooting
This document helps to explain a
little bit about the Telstra Wholesale DSL network, and how the service is
provided to another ISP, such as Pacific Internet, IPrimus, Netspace, Internode
& IInet.
It also aims to help users quickly identify faults on the DSL
network, and how the network integrates into their own.
To start with, I have
divided this document up into the 4 ways that your phone line can be provisioned
for DSL by Telstra.
A Type A Connection is a conventional PPP-based connection
over Ethernet or ATM
About
Telstra
The Telstra-based PPP over Ethernet/ATM looks like
this.
A similar config is used for Pacific Internet / Netspace /
IPrimus DSL connections that go through the Telstra Wholesale Network.
The most
common point of failure within this setup is the link between the IPSN and the
CMUX.
The CMUX is located at your local exchange, whilst the IPSN is also
reffered to as an "Access Concentrator".
There are currently 6 Access
Concentrators use by Pacific Internet on the Telstra network
VET1-EXHIBITIONMelbourne
VWT2-WINDSORMelbourne
NKT1-KENTSydney
NCT2-CHATSWOODSydney
QWT1-WOOLLOONGABBA Brisbane
QCT2-CHARLOTTE Brisbane
Each server has an internal IP
address, starting with 172.31, that linux users will occasionally see, although
this means nothing in the real world as far as gateways and routing goes.
The authentication process relies on many points functioning within the Telstra
Network.
This is where the highest point of failure is. When you authenticate using
PPPoE/A, these are the devices that 'talk' to each other.
|
Your PC/Network
|
Modem/Router
|
Phone Line
|
Exchange/DSLAM/CMUX
|
Access Concentrator/IPSN/Shasta
|
Telstra Radius Proxy
|
ISP Radius Server
|
Rest of the Internet
|
Telstra have provided a definition of some of the terms
used to describe their network.
Telstra authenticates users using a device
called a "Shasta". This is made by a company called Nortel, who have a website about their
product The current software being used is.
V3199:T4:L35:Shasta 5000: iSOS (tm), 2.1.19.3\000 Some people make the mistake of thinking that just because an ISP
resells the Telstra DSL service, they are also using Telstra's overseas and
domestic links. This not the case. In Pacific Internets situation, all data
flows from the telstra network in each state through to a Pacific Internet
router. For example: 1
ADSL.mel.pacific.net.au (210.23.139.201) 32.845 ms 37.389 ms
45.695 ms
2 a6-0-1.mel001.pacific.net.au (210.23.140.85) 27.158 ms
26.011 ms 27.609 ms
3 FastEthernet0-0.mel003.pacific.net.au (210.23.140.162)
32.145 ms 26.874 ms 32.456 ms
4 comindico.gw.pacific.net.au (210.23.140.165) 32.596 ms
27.187 ms 32.416 ms
5 pos2-1.cor01-kent-syd.comindico.com.au (203.194.56.217)
40.774 ms 36.718 ms 39.497 ms
6 ge2-0.cw01-syd.comindico.com.au (203.194.29.249) 38.375
ms 35.295 ms 39.097 ms
7 ip-6.26.194.203.comindico.com.au (203.194.26.6) 41.149
ms 38.711 ms 37.347 ms
8 karl.planetmirror.com (203.16.234.20) 42.275 ms 39.275
ms 38.410 ms
This means that if you are having problems browsing to
certain sites, or sending or recieveing email, the problem is likley to be with
the ISP itself, and have nothing to do with the DSL service Telstra provides. If
you are having a problem logging on the the network at all, then this is when
Telstra may be the cause of the problem. You can find out more about the Telstra
Network from this Interoperabilty
Testing Document. (mirror). This is 46
pages long, and some fairly heavy reading. There is also a similar 'DSL Service
Interface Specification Document' (mirror). The documents are
designed for CPE suplliers, but give a good idea of how the network works. There
is also an unnoficial document written by a Telstra tech which explains some of
their own troubleshooting methods (mirror)
|
A few words about L2TP As of December 2001, Telstra began to allow
ISP's to connect to its wholesale network using L2TP (Layer 2 Tunnelling
Protocol). This change has no visible effect on the way the end user can
connect or authenticate using PPPoE/A, It only effects the way the ISP
connects with telstra on a layer 2 level.
Some
ISP's have claimed that this makes the service more reliable. I cannot
see any reason why this is the case. Data still passes through the telstra
network (on a Telstra phone line, through to an Alcatel CMUX, on to a
Nortel Shasta) where 99% of the current DSL problems lie. The ISP is the
only real party to benefit from these changes. It allows the ISP to offer
"true" dynamic IP's, "boot off" PPP users, and bill timed usage accurately
etc, thus it gives the ISP more control, it does not give the end
user any more reliablity. An excellent summary of the current situation can be found here
I have gathered some relevant infomation relating to how LT2P works. Known L2TP users are
Ozemail and Internode. There are no known conflicts or issues for
end-users who use an ISP that connects using L2TP. | About PPPoE
The majority of DSL connections are PPPoE. If you have an Alcatel ST
Home modem, or a DLink DSL300, you need some PPPoE software, which can be found
here:
When you initiate a PPPoE connection, the
session goes like this.1 - [PADI] -
PPPoE Active Discovery Initiation (client looking for Access Concentraitors)
2 - [PADO] - PPPoE Active Discovery Offer (the nearest AC responds with a
service offer)
3 - [PADR] - PPPoE Active Discovery Request (the client decides on a PADO to
respond to, and does so)
4 - [PADS] - PPPoE Active Discovery Session-confirmation (gives the ok session
initiation)
5 - Session is Started by client (User & Pass)
This means that you should never have to try and
configure or change any settings on the Alcatel Home, or DLink 200/300 yourself.
The entire setup is done on the client software located on your server or PC.
Thanks to the Stallion EPipe debugger, we can see the PPPoE
authentication process in
action. This log refers to a failed authentication attmpts, but it shows us
enough to see how PPPoE works. Also of intrest here is the CHAP reference to
"ppp@shastanets.com" which is the default name/setting for ppp terminations at
the Nortel Shasta
A correct PPPoE implementation can be seen in RFC2516
About PPPoA
PPPoA stands for PPP over ATM. I'm not
going to start to get into ATM or various other protocols, as there is no
visible difference to the end user. Most small basic routers (such as the
Alcatel ST Pro, or the DLink 500) implement PPPoA. An internal (ISA or PCI) or
external (USB) adapter provides the user (transparently) with an ATM interface
(as opposed to an ethernet one for NIC/PPPoE). Due to the low cost of these
devices, they are distributed or sold with many of the cheaper residential DSL
packages ( Pacific Internet or
Internode). Thus, many end-users
will end up using PPPoA for the DSL connection.
Modems that need to
use PPPoA:
D-Link DSL 200 USB Modem - Being used by Pacific Internet HomeDSL
customers, and we have some information on how
to set this up for a generic Windows-based installation. The USB drivers
come on a CD with the modem.
Alcatel SpeedTouch USB. Currently being distributed by
Netspace, for residential DSL. It has a very unique design,
and is currently the only product on the market with Linux USB drivers.
In all instances, a PPP connection is made in the normal method of
the operating system such as Windows' Dial Up Networking (DUN). No configuration
should be done on the device itself. All that is required is to install the
specific software drivers on the operation system of the PC. This sort of
connection is not recommended for low-spec machines, due to the fact that they
will draw power and resources from an already-drained host machine, and possibly
interfere with the DSL connection.
Hardware Issues with D-Link 200 USB
modems. The D-Link 200 has some unconfirmed power management issues when
used with some VIA motherboards. (inparticular, variants of the VIA
VT8366/VT8233 chipset). The modem intermittently 'drops dead' with no lights,
and no activity, a reinstall fails to find the modem, or the modem won't
install, or be detected at all, or will be detected incorrectly). We believe the
problem is some of the latest VIA USB controllers, they don't seem to be able to
manage the resources required for a D-Link 200 USB modem. Its well known that
the USB modems can use up to 80% of the bandwidth of a USB controller, but no
other motherboard seems to have the problem.
Workarounds for this
problem are
ADSLGuide,
UK - Finding USB Bus Bandwidth D-Link 200 FAQ. -
Why DSL-200 stops responding after leaving it on for a while? Microsoft
KB - Windows XP Does Not Detect Your New USB Device And some other issues
are mentioned in this rough
document (Thanks Polly M)
PCI-DSL Modem cards
Generic PCI
DSL modem-cards are made by d-link and alcatel, and are the cheapest DSL modem
overall. The fact that you have to open up your PC to install them seems to have
limited their popularity a little. The only people I know that use PCI modems
use them on Debian/GNU Linux boxes. The
modem comes with a generic driver that provides a ethernet-style bridge
(referenced as BRI1), and PPPoE is then used to terminate the
connection.
Router information:
The following devices are
known to use PPPoA to terminate a DSL connection.
Alcatel Speed Touch Pro - All setup information can be found here.
Cisco 827 - Intructions and a sample
config - Done with Pacific Internet values, so you may need to substitute
these. (Thanks Lalor).
Dlink DSL 500 - A correct PPPoA setup looks like this. The DSL500
is very configurable, and there are lots of nice things you can do with it.
For more info see D-Link's support pages.
The DSL-504 is a device with the same firmware, but with 4 extra ethernet
ports, so it can act as a hub. (nice).
Netgear rt314 - A virtual web
setup has been created, use the "Wizard setup" to configure the DSL
account.
Netopia r6100 DSL Router - Support document
by Paul Alexander
Nokia m1122 - Thanks to the people at iPrimus, who supplied this pppoa config for the
Nokia dsl modem/router. Some problems have been noted with Alcatels
PPPoA implentation that appears to prohibit running a VPN through the PPPoA
interface. I believe its something to do with the way the fort fowarding is set
up. Despite many attempts, I have never seen this actually work, and would love
somone to prove me wrong. There is also a Bigpond Direct document which has some
interesting sample configs for various routers that may be of interest (mirror)
PPPoA is
referenced in RFC2364. More
PPPoA information is at Everything2.net
(used as a reference).
Troubleshooting : No
Linesync
On an Alcatel ST Home modem, there is a
light called the "linesync", which is located second from the right-hand side.
If this light is flashing, it means that the modem cannot locate the CMUX at the
Telstra End of your DSL Line. A linesync issue is universal problem, regardless
of how the DSL is configured on your phone line
Possible Issues
Phone Line - The first thing to check is to see if there is a dialtone on
the phone line. Check cables, plugs, sockets, filters, other devices. This
type of fault is usually hardware related. If you cannot hear a dialtone then
you can speak with the Telstra faults center directly on 13 2999
Telstra Fault (a) Widespread - If a CMUX on the Telstra network is faulty,
it will usually also affect other customers in the surrounding area. These
faults are indiscriminate, widespread, but are usually rectified by Telstra
within 6 hours.
Telstra Fault (b) Isolated Fault / Incorrect Provisioning - Other faults
can be a little more obscure, and I have seen many caused by human error
relating to Telstra Technicians. Telstra will also de-provision the DSL
service from your phone line if you need to move it, change services, change
billing details, or ownership of the line.
Encapsulation Type - If your modem is intermittently losing sync, or not
syncing up at all (whilst other equipment will) check the line encapsulation
type on the modem. The correct setting should be "Multimode" or "ANSI" (this
is more or less an auto-detected setting) - or you may wish to set your modem
to use "G.DMT", which is what Telstra uses. In Short : A linesync
issue is almost definitely a Telstra Issue, as your ISP usually does not have
access to this part of the network. It is also something that can effect any DSL
line, regardless of if it is provisioned as a type A, B, C or D
Troubleshooting : No Data on Authentication (no
comms)
If your linesync is ok, but your PPPoE client reports a
'session-timout' or 'Time out while waiting for PADO' whilst logging on. We are
encountering a different type of fault. It usually means the CMUX is ok, but the
connection between the CMUX and the IPSN is not.
There are 3 types of
faults here:
Check all the network cards involved, as well as any hubs, cables, other
devices etc. Most NIC's have a green LCD light on them, that should be
statically on to indicate the ethernet connection is working on a hardware
level.
Widespread Telstra Outage - Failure of Access Concentrator. If this is
beeing rebooted, or under maintenece, then usually the majority connections
across your entire city/state will also fail. Because this will affect a vast
ammount of clients, Telstra and associated ISP's usally post notage of this,
and you may be able see this information on their tech websites (Telstra & Pacific) Such faults usually last
for under 30minutes, and the worst example of this was one that lasted for 6
hours.
Single Telstra Fault - Data Lockup. This mostly affects PPP-based
customers. Sometimes, for no good reason, the 'data tunnel' between the CMUX
and the IPSN will corrupt itself. The 'tunnel' is how they refer to the PVC
that forwards all your traffic from your CMUX at your local exchange out to
the IPSN. Telstra's solution to this ongoing problem is just to break and
re-establish this link, generating a new PVC from the IPSN to the CMUX. Your
ISP will have access to your PVC identifier, and this will change when the
'tunnel reset' happens. If your connection 'locks up' when there are no known
outages on the network, your ISP should request a rebuild straight away. These
may take around 24 hours for Telstra to perform. (acronym overload?)
So
why do these happen? The best answer I have seen to this was provided by a
current telstra helpdesk staffer in a series of emails, and fit in with the
what most ADSL support people are observing when dealing with Telstra
| PPPoE is a dodgy protocol, broadcast protocols running through
it it, like udp and such, can cause the ppp session to lockup and
reset. This disconnection of the pppoe session should not however
cause a problem with the tunnel. Blame nortel, or alcatel, or telstra,
or all three, maybe even the easter bunny as well. Probably nortel
though, the atm pvcs lack of traffic flow is most likely caused from
the IPSN end of thing, as it provides you the ppp session. |
- Borderline attenuation problems (usually with downstream),
this also accounts for customers going in and out of sync alot and
that causing ppp sessions to die, this repeatedly disconnected thing
sees a higher frequency of the tunnel dying.
- cmux problems,
don't know the cause, [Telstra] will sometimes initiate the software
being reloaded onto the cmux, sometimes the cmux port is swapped for
customers that get alot of tunnel problems. But pretty much telstra
have stopped blaming alcatel for the tunnel problems, it is still
under 'investigation' by Nortel and Telstra technicians.
- as
i was saying with udp based things, saturating upstream with these
type of things, when your machine is trying to send out to many
packets, that really increases the likely hood the ppp session will
die and the tunnel too. for example, 512/128 customer, running a
quake3 server ( I don't know why, but they do) - 4 other people
connect in with rates of 5000+ - bam.
- PPPoE clients, seems to
happen with all pppoe clients, however, in regards to things like
broadcast protocols saturating your bandwidth, we tend to see OSs with
better tcp stacks handle things better. Linux/rp-pppoe tends to get
alot less of these types of issues, where as win98/ME are culprits.
Probably the send-q handling, I dunno really.
| This also raises the
possibilty that these data lockups could somehow be 'engineered' as a form of
Denial of Service attack against a particular user, although I have never seen
any evidence of this. Another application that is known to cause problems with
PPPoE is the Microsoft "Netmeeting" program. In Short: It would seem
that this part of the Telstra network/infrastructure is inherantly faulty at
this level. There has been some some interesting discussion
and observations made already with regards to the Nortel Shasta and its
capabilties in handling a large DSL Network. Officially, Telstra have never even
acknowledged that this issue exists, and this would seem to be the first step
involved in solving the issue, to avoid a general migration away from their
wholesale network by their largest customers.
Troubleshooting :
Other Issues
Slow Speeds - These are not usually DSL related, (I have never had any
capacity issues on my line) however, and easy way to test is to do an FTP
transfer from your local ISP's mirror server (Pacific) On an stand-alone connection
you should be able to get these speeds
256/64 = above 20KBytes p/s
512/128 = above 40KBytes p/s
1500/256 = above 120KBytes p/s Your ISP can investigate if
Telstra have provisioned the line at the correct speed (or some modems will
tell you) or the line is running at a high capacity. Telstra can see if the
DSL service is taking up more than 50% of total line usage. This is usally
tested before the line is provisioned to avoid this situation
MTU (Maximum Transmission Unit) related issues (should) only occur when
using PPPoE. Common symptoms of this is very slow performance, problems
FTP'ing and some webpages just not coming up for no particular reason. The
reason these issues occur is to do with an inherent flaw in the way PPPoE
works as a protocol Fenn has written
about how these problems happen, and how to work around them. This document from DSL reports
provides some tests you can do to see if MTU is the cause of these
problems
Alcatel Modems have some
terrible security flaws (mirror)
Be sure to check sites like Whirlpool, CableGuy & Ausforum for current discussion about new
troubleshooting, problems, tricks & tweaks.
Issues like speed and constant dropouts can be caused by the DSL, but they
are just as likley to have nothing to do with it. Be sure to check everything
- phone lines, extention cords, protocols, modems, sockets, plugs, dust,
switches, software, hardware, your operating system, etc. Your ISP will only
speak to telstra about the issue when they are sure that its not anything else
thats causing the problem.
[top] A Type B Connection is a pure Bridged Ethernet connection
between you and Telstra
This is the type of connection I enjoy
using at home. In my opinion its the most simple, reliable and trouble-free DSL
connection available (Its also slghtly fast as there is no PPPoE involved).
Pacific Internet do these type of connections for some corporate customers. Most
ISP's will not do this for a residential account because of the 2 IP address's
that are wasted making the connection.
A typical TypeB connection looks
like this:

The configuration for this is
done on your NIC. The settings are:
W2K : Start => Control Panel =>
Network & DUconnections => rightclick = Properties => TCP
IP Address : One from your allocated network
Default Gateway : Telstra End GW
NetMask : Your allocated Netmask (eg 255.255.255.252 for 2 usable address's)
DNS : Your ISP's settings
These settings should be given to you by your
ISP.
My own connection was done on a RedHat Linux box, and I have
documented the results. The
good thing about these connections is that they do not suffer from 'data
lockups' like PPP connections do, as there is no authentication process. I have
experienced 99.99% uptime with this configuration.
Troubleshooting : No Linesync
A linesync issue is the same with any type of DSL-enabled phone line,
regardless of how this line is provisioned. The tips of have already
outlined will apply here also.Troubleshooting : No
Data Transfer (no comms)
Check all the network cards involved, make sure each NIC is responding.
From the machine directly connected to the DSL modem, you should be able to
ping your Telstra gateway address (and anything else). if this request "times
out" it could be an indication that Telstra's Access Concentrator is offline
completley. Your ISP is the only other party that has the abilty to ping that
gateway, and this is the first thing a helpdesk rep will check.
If you can ping the telstra gateway, but recieve networking-based errors
on other external IP address, then we need to double check your routing table
and your ethernet interface configuration. The command is 'route print' (you
should see a default gateway here, with the IP address of 0.0.0.0) and
ipconfig (ifconfig on a linux box) which will show the gateway for each
interface. My 'ifconfig' looks like thiseth0
Link encap:Ethernet HWaddr 00:10:4B:79:C4:30
inet
addr:210.23.131.202 Bcast:210.23.131.255 Mask:255.255.255.252
UP BROADCAST RUNNING
MULTICAST MTU:1500 Metric:1
RX packets:79966
errors:0 dropped:0 overruns:0 frame:0
TX packets:77873
errors:0 dropped:0 overruns:0 carrier:0
collisions:1
txqueuelen:100
RX bytes:45481255
(43.3 Mb) TX bytes:13608460 (12.9 Mb)
Interrupt:10 Base
address:0xde00
And my route table looks like this. (ADSL.mel is the
Telstra gateway, default is the same as 0.0.0.0)
Kernel
IP routing table
Destination
Gateway
Genmask Flags Metric
Ref Use Iface
210.23.131.200
*
255.255.255.252 U 0
0 0 eth0
127.0.0.0
*
255.0.0.0 U
0 0
0 lo
default ADSL.mel.pacifi
0.0.0.0 UG
0 0
0 eth0
In Short : we need to ensure that a) You PC can
identify your network interface. (hardware detection) b) Your
network interface can idenitify itself, and the network that it is apart of.
(interface config) c) You network needs to have a place where it can
send data to other networks. (routing table)
This type of
conection forms the base of a Type
D configuration. A Type D will attempt to build on your current setup, using
your current bridged ethernet to route another network down the DSL line.
[top] A Type C Connection is a way of routing networks down a DSL
line using a router device
These connections must be made using a
DSL modem/router. The have been tried and tested on an Alcatel SpeedTouch Pro
only, although they should work on most Cisco routers and the D-Link DSL 500.
A routed connection looks like this.
On your router, you need to:
(a) Configure your WAN interface with your User-End IP address,
and set a gateway as your Telstra-GW. Set an appropriate subnet mask
(Pacific Internet allocate a /30 for these connections, so a subnet mask
would be 255.255.255.252). From this point you should be able to ping the
Telstra Gateway without any hassle
(b) Configure the route
for your network, as your network will travel "through" the frame we have
just created.
(c) Configure the router with another IP
address, taken from your network. (Most people choose the lowest
ip)
(d) Allocate another IP address to a client machine on
your network, and set the gateway as the network IP you have just given
your router.
(e) Make sure the router can see both subnets.
From your router you will be able to your Telstra GW, and your client
machines should be able to see the rest of the internet
|
This style of configuration differs between
routers, as do the troubleshooting methods. At the moment 3 modem/router devices
are supported. Aside from what I have mentioned here, I have been told by third
partied that the Netopia r1600 will work with this configuration, as will the
iprimus-supplied nokia's.
Alcatel Speed Touch Pro - I have written explicit documentation on how
to set this up.
Cisco827 - Command setup and running
config
D-Link DSL500 - The out-of-the-box 500 series does not come with
full cip/ipoa support. The firmware inside the modem needs to be upgraded. The
firmware was only released in January 2002, and it can be found here. The firmware is
labelled"DSL-500R201b5AU.exe". There is also firmware available for the 504,
but from what I am told, the only real difference is in the labelling.
Troubleshooting : General
See here for 'no
linesync' troubleshooting
To establish that there is a 'no data' issue, attmpeting to ping your
Telstra gateway from your router. If it does not return a ping then it could
mean that Telstra's end is completley offline. At this point it may be a good
time to ring your ISP to get confirmation.
This following advice comes from Lee DeSouza of Pacific Internet's DSL
provisioning department
For a new (or modified) connection, easiest way to tell if it's
misconfigured (or rather, a dead giveaway) in terms of mode (IP over ATM
vs Ethernet) is to ping the customer and get them to watch the line rx /
adsl act light... if it appears to blink once a second (ie: in time with
a constant ping once per second) but you get no reply - it's probably
provisioned wrong (or of course, the other issue would be the cust not
correctly setting routes).
Naturally, for a customer requiring a
config D that actually (incorrectly) has a C, the access concentrator
sends ATM-encapsulated cells to the end-user, to which the Ethernet
device on the customer end has no clue what to do with. And vice versa.
|
The only other real point of failure
is the router itself. I have seen that the SpeedTouch Pro needs to be rebooted
on occasion, for no apparent reason. The first thing to check in any instance is
the route table - make sure its the same as when you left it.
The reason
that there is not much documentation for this type of connection is because the
user has full control over what is done with his or her network. All telstra do
is provide a gateway for you to send data to, the rest of the configuration is
done on your router. For further troubleshooting with the ST Pro, I suggest you
read my config document
[top] A
Type D Connection uses bridged ethernet to route additional networks on a DSL
line
The first thing to understand is that a Type D
configuration is an extention of a Type B. With a Type B, a fairly
straightfoward Bridged Ethernet connection is established. A Type D uses this
connection to route more networks down the DSL line, in a similar fasion ot a
Type C.
You can utilise this connection with any PC that has more than 1
network interface. It looks like this.
You should have been
allocated your End-User address and gateway address by your ISP. You will need
to configure the network card appropratley, and then ping the GW end. Assuming
that you get a response from the gateway, you can proceed to configuring the
second network card with an IP address from your allocated range.
A
customer set this up on his debian box, and emailed me to tell me how he did it
Troubleshooting : Networks
If the linesync light is flashing, see here.
Before reading anything else, get an idea of how to do some troubleshooting for the Bridged
Ethernet This has been taken from the Type B Config, as we need to
make sure the same bridged ethernet configuration is working before we move on
to testing the routed config. Some ISP's allocate internel IP ranges to make
up this first part, and you shouldnt need to treat these any differently to
public IP space.
As with all network connections, check all the network cards involved, as
well as any hubs, cables, other devices etc. Most NIC's have a green LCD light
on them, that should be statically on to indicate the ethernet connection is
working on a hardware level.
Now see if you can ping the telstra-end gateway given to you by your isp.
If this is down, then their is a high possibility that the Telstra IPSN is
completley offline. If this is a new connection make sure you clarify the
gateway details with your ISP - its also possible that Telstra have configured
this incorrectly
Your ISP can run a test
to make sure the line is provisioned with a "Type D" and not a "Type C" or
other configuration.
Provided you can ping the GW successfully, but are having problems pinging
elsewere, the routing table is most likley going to tell you whats wrong with
the connection. I have drawn up a standard one here, as seen from the server
machine. It assumes that 203.10.20.0/24 is your network, 210.23.131.200/30 is
the bridged ethernet. (210.23.131.201 is the Telstra GW) The command is 'route
print', (from the CLI) and this table format may vary slightly between
windows/linux/bsd.
Destination
Gateway
Genmask Flags Metric
Ref Use Iface
210.23.131.200
*
255.255.255.252 U 0
0 0 eth0
203.10.20.0
*
255.255.255.0 U 0
0 0 eth1
127.0.0.0
*
255.0.0.0 U
0 0
0 lo
0.0.0.0 210.23.131.201
0.0.0.0 UG
0 0
0 eth0
Interface eth0 should have the IP address of
210.23.131.202 (network card connected to DSL modem) Interface eth1 should
have the IP address of 203.10.20.1 (network card connected to your LAN)
The routing table on a client machine is a little more simple. It also
assumes that you choosen given the lowest IP on your network to be second NIC
on your server, and this is the gateway for the client machines.Destination
Gateway
Genmask Flags Metric
Ref Use Iface
203.10.20.0
*
255.255.255.0 U 0
0 0 eth0
127.0.0.0
*
255.0.0.0 U
0 0
0 lo
0.0.0.0 203.10.20.1
0.0.0.0 UG
0 0
0 eth0
As with a Type B, we need to ensure that a) You PC
can identify your network interfaces. (hardware detection) b) Your
network interface(s) can idenitify themseves, and the network that they are
apart of. (interface config) c) You network needs to have a place
where it can send data to other networks. (routing table
config)
Once these connections are set up correctly, I have
never seen them change dramatically, to the point where they need serious
troubleshooting. This is simple ment to give you an idea of how the setup should
work.
[top] Other Links and resources of use.
The Cable Guy - Excellent
resource site for tweaks / troubleshooting, tips and problemsolving for all
types of broadband connections.
Whirlpool - Broadband News Site -
also check out the forums for discussion and ISP reviews. Also home to the Australian Broadband FAQ
BroadBand Choice - Lists
australian ISP's, shows benfits, of each ISP compared
AusForum - Broadband Users
discussion site, topics range from DSL, Linux, Cable, Gaming and everything
else. Very busy site.
aus.net.access -
Newsgroup based around ISP discussion and access problems.
Becsta's Linux-ADSL guide -
Informative Linux/DSL/Bigpond FAQ
DSL Reports - US-based site with
the latest DSL tricks and tweaks
ChasmS - Contains virtuals machines
for remote DSL troubleshooting
Greystorms DSL Page
- For using ICS software to share your DSL connection
adsl.cutw.net - My own DSL-info
dumping ground
Pacific Internet Software
Mirror - Where I get all my software from locally Definition of Terms & End Notes.
I
gathered some of these terms from this document.
A complete definition of terms can be found here ADSL
= Asymmetric Digital Subscriber Line
ATM = Asynchronous Transfer Mode
CMUX = Customer Multiplexer
CPE = Customer Premesis Equipment
DAC = Digital to Analog Converter
DSLAM = DSL Access Multiplexer
IPSN = Internet Protocol Services Node (also known as TAS / Access
Concentrator)
LT2P = Layer 2 Tunneling Protocol
MTU = Maximum Transmission Unit
NIC = Network Interface Card
PPPoA = Point-to-Point Protocol over ATM
PPPoE = Point-to-Point Protocol over Ethernet
PSTN = Public Switched Telephone Service
PVC = Permenent Virtual Circuit
Disclaimer
This Document is to help you troubleshoot your own DSL connection, but it does
not come with any warranty or liablity. It is yours to play with. Any opinions
conveyed in this document are mine, and not that of any other persons,
including my employer. This document may be freely distibuted so long as it is
not used for profit, or modified in any way.
Credits
This document was written by myself, using the knowedge gained from working at
Pacific Internet (Australia). This document has ended up being more detailed
and complex than what I first set out, so most lilkey I will be writing an
abridged, less-technical version at some stage. This document would not be
here without the help of people like Byron Brink and Fenn Bailey, and the
technical staff at Pacific Internet. I am also greatfull for the help and
information contributed by some unnamed techincal staff at Internode, Telstra,
and iPrimus.
Revision 1.3 12/06/02 | Changelog | Access Logs
Please send errors / questions to james [at] pacific.net.au
(c) James Mollison 2001-2002
[top]
|