|
ISDN
Trouble-shooting
1. DETERMINE
WHAT THE PROBLEM IS:
More than likely you already know that you have no Internet access.
The first step is to figure out where the connection is dying out.
The best process for doing so is by process of elimination.
First, if you have no access to email or web sites, power cycle
the unit. This means; turn the router off, leave it off for roughly
20 seconds; unplug the ISDN line from the back of it; plug the ISDN
line back in; turn the router back on. When the router is turned
back on it will reset the ISDN line and restore the system memory
resources. When the test light stops flashing, you should have a
power light, D light (symbolizes signal from ISDN line), and a link
light. When you see this, open the browser and attempt to find a
web page.
2. NO D-CHANNEL
LIGHT: If the D light on the router does not come on after the
test, it means that no signal was picked up. If this happens, turn
off the router for at least a minute and turn it back on. If the
D light stills sits unlit, unplug the ISDN line from the router
and dial the SPIDS. The SPIDS are the actual phone numbers that
identify the ISDN circuit with your phone line provider. If the
ISDN line is configured for Voice and Data, the phone should ring
through. If the Data/Voice ISDN lines respond busy, then there is
a problem with the line. If you simply have data only
ISDN lines, you will get a normal busy signal. If you get an operator
message with either line type or the D-Channel light is off you
should call your phone carrier and open a trouble ticket.
3. D-CHANNEL
LIGHTS UP: With the D light on, try accessing the web. The router
should dial up and then the web page should be available. If no
site comes up, check to see if B1 and B2 lights are on. B1 and B2
show that the ISDN line is in use represents both channels for the
ISDN line. If the lights are on but you are still unable to transfer
date try pinging our DNS (64.7.161.12). If no reply, try a trace
route to our DNS server and find where the connection fails.
Pinging
the router: Make sure that workstation can ping the router
on the network. To ping the router on the network, first you will
need to find the IP address of the router. You can do this by
clicking Start --> Run --> [type] Command
[hit Enter]. At the DOS prompt [type] winipcfg (for win9x)
or [type] ipconfig (for winNT, Win2k, winME, winXP). Find
the Gateway Address from the information that appears
and try and run the ping command below. To do this, click Start
--> Run --> [type] Command [hit Enter].
From there,
[type]: ping [IP
Address/domain of the destination you are trying to reach]. (Example:
ping 192.168.0.1). If the user is unable to ping the router
there may be a dead cable or hub. There also might be a server
that needs to be rebooted. If you have a more complex LAN (Local
Area Network) setup, there is another useful command that helps
to pinpoint exactly where the connection is dying. It is called
Trace Route and is similar
to a ping in how it transmits a packet to a destination address.
But, a Trace Route illustrates every jump that the data packet
takes before reaching the destination address. This command can
be executed by clicking Start --> Run -->
[type] Command [hit Enter]. Once at the command prompt
[type] tracert [IP Address/domain of destination you are
trying to contact]. (Example: tracert 192.168.0.1). Use
a destination IP address outside of your network. Using 64.7.161.12
(our Network DNS server) will be sufficient. See Figure A
below for an example of a successfully traced route to www.yahoo.com.
If you are not able to ping or tracert to any destination address
beyond your computer, the problem is probably due to incorrect
network protocol configuration.
Figure A:

4. NETWORK
CONTROL PANEL SETUP:
Depending on the network setup, the Network Panel should have either
one of two configurations: DHCP or statically assigned addresses.
If a server or the router is running DHCP then the Network Panel
will be; Server Assigned IP Address, Use DHCP, No Gateway installed,
Disabled DNS. This means that the computer is supposed to be assigned
an IP address, Gateway, and DNS. To check if the computer is picking
up correct information run 'winipcfg' from the 'Run' menu (only
works on Win9x). On a WinNT or 2000 machine, run 'ipconfig' from
the command prompt to check out the addresses.
5. IF CONNECTED
THROUGH ONE CHANNEL ONLY: You will want to contact our Network
Technical Support to check the integrity of your router configuration.
If you have the password to a router that our Network supplied
(probably a Netgear 338) than you will simply want to log into that
router and verify that the SPID numbers for your ISDN lines are
configured correctly. The same goes for if you purchased your own
router. If the SPID numbers are entered into the router correctly,
there is more than likely a problem with one channel on the ISDN
line. In which case you would want to contact the provider of those
lines for assistance.
6. SPORADIC
PERIODS OF CONNECTION PROBLEMS: If you are intermittently having
problems getting to websites and/or e-mail, try and log when problems
occur. If they are predictable, i.e. happen during certain times
of the day or, if the problem is totally sporadic. This could be
caused by blatant over-utilization of available bandwidth or, a
loss in ISDN signal. To check if it is a line problem, monitor the
D-channel light when the outage occurs to see if it stays lit or
dies. If it dies, you would want to contact your telephone company.
If it seems to be a possible over use issue and the problems seem
predictable follow the steps below.
The first
thing you will want to do is run a constant ping
on the router. This will help monitor the traffic to the Internet.
Open the command screen (Start --> Run --> command [hit
enter]). Type ping -t [router IP address] and hit Enter. The -t
will force a continuous ping on the destination address specified,
until Ctrl + C is pressed on the keyboard to stop the command.
If the connection
problems are currently happening, watch the number of dropped
packets (request timed out) and the number of returned packets.
Watch the time in milliseconds (MS) that it takes for the returned
ping requests (see figure A below). If the ping time increases
sharply in a matter of 6 - 10 pings and then a request timeout
appears, the culprit is most likely related to heavy usage on
the network. This means clients within the LAN are using all their
allotted bandwidth, and as a result, packets are dropped. Therein
lies the cause of their sporadic connection losses.
Figure
A:
Reply
from [IP address]: bytes=32 time=24ms
Reply from [IP address]: bytes=32 time=100ms
Reply from [IP address]: bytes=32 time=156ms
Reply from [IP address]: bytes=32 time=240ms
Reply from [IP address]: bytes=32 time=550ms
Reply from [IP address]: bytes=32 time=1200ms
Request timed out
Request timed out
Request timed out
Reply from [IP address]: bytes=32 time=24ms
Reply from [IP address]: bytes=32 time=200ms... |
If the periods
of outage are totally unpredictable the most common cause
is the ISDN lines or the router configuration. Follow the steps
above and do a continuous ping on the router.
If the connection
problems are currently happening, watch the number of dropped
packets (request timed out) and the number of returned packets.
Watch the time in milliseconds (ms) that it takes for the returned
pings to reach us (see figure A below). If you see that the request
time out appears randomly with little or no increase in
the ping times (see figure B), then proceed to trouble-shoot the
connection starting on step 1.
Figure
B:
Reply
from [IP address]: bytes=32 time=24ms
Reply from [IP address]: bytes=32 time=13ms
Request timed out
Reply from [IP address]: bytes=32 time=20ms
Request timed out
Reply from [IP address]: bytes=32 time=24ms
Reply from [IP address]: bytes=32 time=40ms
Reply from [IP address]: bytes=32 time=25ms
Request timed out
Reply from [IP address]: bytes=32 time=13ms
Request timed out
Request timed out
Reply from [IP address]: bytes=32 time=26ms... |
IPConfig
/ WinIPCfg Explanations
The above DOS
command is used to find LAN information such as:
Connection-specific
DNS suffix - This is the specified domain of the resolving DNS
server for a named network.
IP address
- This is the address the computer is either statically assigned
with, or the (Internal/Real) address the DHCP Server assigned it
with upon initial LAN access.
Subnet Mask
- This is the specified Subnet assigned to the local host which
determines the size and/or class of network for the associated client
IP address.
Default Gateway
- This is the gateway IP address. This is typically the last
hop a client IP address takes before reaching the outside Internet.
If a DHCP/NAT server is running it will probably be the address
of that server. If not, then a router IP address will probably be
the assigned gateway address.
The commands ipconfig and winipcfg are the identical
commands for different versions of Windows. The ipconfig command
works on winME/2000/NT/XP. The winipcfg command is only used on
win95/98. These commands are especially useful when trying to locate
and test connectivity problems for dedicated customers or, customers
running a Local Area Network. The information from this command
provides you with the ability to determine whether the problem is
related to internal IP/Subnet designation, or, with the gateway
router/server. If the IP address is associated with an incorrect
subnet, or, is on a completely different network as the other machines
and router, it will be visible when running this command. This is
also useful for when customers are unsure of the router address
and configuration review/changes are necessary. The router will
typically be the Gateway address listed by this command (but, can
sometimes be a server).
Ping
Explanations
The ping
command is used to determine whether a specific address is available
and properly responding to packet requests across a LAN/WAN. The
ping command will respond with the following information:
"Reply
from xxx.xxx.xxx.xxx: bytes = ## time<##ms TTL=### "
Reply -
This is a response given from an IP address (the from address listed)
that is correctly responding to a packet request. You will typically
receive four consecutive replies that should all respond with relatively
the same statistics. You can change the number of replies from four
to continuous by including a ' -t ' after the first portion of the
command. To stop pinging the IP address, simply hold the Control
key down on your keyboard (Ctrl) and press the "C" key.
See example below:

Request timed
out - If a ping reply does not make it to the destination address,
or fails to receive a timely (if any) response from destination
address ("timing out"), it is safe to assume that there
is a problem with the equipment configuration. The problem may also
be a Layer 1 issue, which would include router hardware failure
and all lines and connections between the ISP and the destination
host (Repeaters/Smart Jack/ISDN/DSL/T1 lines/etc). See example below:

Destination
host unreachable - If a ping reply responds with a "host
unreachable" type error, it is probably due to some form of
firewall or port block set on the gateway router or, before any
destination address that you are trying to reach (be it a computer
within the network). It could also be related to any form of routing
issue (see trace route explanation). See
example below:

Bytes -
This is the amount of bytes sent in the ping packet (and can be
changed for bandwidth/speed testing purposes). To change the amount
of packets sent type a '-l' and the amount of Bytes you want the
packet to include after the first portion of the ping command. See
example below:

Time -
This is the amount of time in milliseconds that the ping responded
with. This is a "round-trip" time that the packet of information
had traveled in. This information is important in situations when
dedicated service customers report intermittent connection problems.
Assuming that all other bases have been checked, you could run a
constant ping to the your gateway address (typically a router) and
have the ability to monitor traffic and ping times. If serious jumps
in ping times are noticeable in situations where only a single computer
is browsing, there is presumably some form of throughput issue causing
the problems.
Trace
Route Explanations
The trace
route command is similar to a ping
command. But, rather than simply checking to see if a specified
address is responding, the command traces the exact path a packet
of information would take to reach its host. This is especially
useful for trying to diagnose routing issues and wide network outages.
It allows a user to see certain hops to pin-point the
exact location of where a connection has actually failed. Another
obvious difference with trace route is that it does displays both
IP address and Host name for each hop present in a packet
destination path. As opposed to ping, which only does
forward or reverse DNS depending on whether you ping a host address
or an IP address.
|