Under the new version of redhat I have noticed that the ping component
appears to be updated. However, as part of the update, if the -c count
flag is used ping will not stop until "count" pings have been returned.
The documentation reports that if ping goes "count" counts without a
single return it will also abort. However, when testing my network, if a
server is down and ping cannot reach it, it will continue to ping until
manually stopped, no matter what "count" is specified. I am running ping
inside another program that automatically checks the network status every
fifteen minutes so there is no way anyone can watch that program to make
sure that it completes every fifteen minutes. Is there a fix for the ping
command to abort it if no response is recieved after "count" tries?
Red Hat 6.x does not use the ping from netkit-base, but rather a version
from Alexei Kuznetsoff that is more appropriate for modern linux kernels.
That version has changed the semantics of "-c N" to wait for N received
packets, rather than N sent packets. This change is necessary to acquire
rtt with constant statistical accuracy, the other use of the "-c N" option.
Use the new "-w secs" flag if you want a deterministic exit for hosts that
may not be reachable. Or, recompile the netkit-base (now called "inetd") with
the netkit ping.