3.3

Network Failure
103
In the previous ping command examples, the number of ping requests sent would continue
endlessly. To control the number of ping requests use a specific number to limit the number
of requests. For instance, in the following examples, five ping requests will be attempted as
noted by including the number 5 in the ping command.
NRCS-A$ ping 5 152.165.17.110
NRCS-A$ ping 5 nrcs-a
In the case of an iNEWS Workstation that is having problems connecting, the first step is to
try pinging the server from the workstation to make sure they can see each other on the
network. (The syntax for ping on Windows is different than the Linux ping command.)
If computers are unable to ping each other, check for a loose or disconnected network cable
or a hub that might have lost power.
netstat -i Command
One diagnostic command you can run on the server to quantify network errors is the
netstat -i command. The netstat -i will show:
• How many packets have been transmitted
• How many input errors (RX-ERRs) have been detected
• How many output errors (TX-ERRs) have occurred
Output Errors (TX-ERRs)
The primary barometer of network health is the output errors (TX-ERRs) column. The
computer will try to transmit a packet 15 times before chalking the attempt up as an
TX-ERR. It will then try to transmit again 15 times and may wind up incrementing the
TX-ERR count. Clean networks will show zero TX-ERRs, or no more than a few.
Of more concern than the raw number of TX-ERRs is how quickly they are increasing. If
you are picking up an TX-ERR every minute, this would be indicative of network problems
and a disconnect may be imminent.
Periodically run a netstat -i so you can get a baseline feel for how many TX-ERRs your
system produces each week or month.
Input Errors (RX-ERRs)
Input errors (RX-ERRs) are fragments of packets or unrecognizable packets. Systems
co-existing on a Novell Local Area Network (LAN) generally show many RX-ERRs,
although they do not seem to cause problems.