Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 11 Jan 2014 14:32:41 +0100
From:      Mark Martinec <Mark.Martinec+freebsd@ijs.si>
To:        freebsd-net@freebsd.org
Subject:   Re: Merge ping+ping6 and traceroue+traceroute6 to single =?UTF-8?Q?utilities=3F?=
Message-ID:  <a46d4e716d15f48b1d4bd32c22dce45c@mailbox.ijs.si>
In-Reply-To: <52D14140.3090003@gibfest.dk>
References:  <1063008459.20140111160525@serebryakov.spb.ru> <52D14140.3090003@gibfest.dk>

next in thread | previous in thread | raw e-mail | index | archive | help
> On 11-01-2014 13:05, Lev Serebryakov wrote:
>> Is here any project to merge ping/ping6 into ping and
>> traceroute/traceroute6 into treaceroute? As IPv6 becomes more common 
>> these
>> days, it is very inconvenient to have these utilities separated.
>> 

2014-01-11 Thomas Steen Rasmussen wrote:
> I hope not, these should remain seperate, allow me to explain:
> 
> There is a good reason these utilities are seperated into v4 and
> v6 specific versions, while other tools support both. The reason
> is that ping and traceroute are network troubleshooting utilities
> that are only used for verifying/testing network connectivity.
> 
> When testing network connectivity you are usually thinking about a
> specific protocol. Having seperate versions of the tools removes the
> ambiguity for hostnames with both A and AAAA records. If you want
> to test v4, use ping, if you want to test v6, use ping6.
> 
> Normal network enabled utilities like telnet or ftp or nc support
> both because when using those you usually don't care about the
> address family used, you just want to connect. This is a significant
> difference from using ping or traceroute where you almost always
> want a specific address family, depending on what you are testing.

While the argument may be valid from some particular point of view,
I'd be very much in favour of having a unified utility.

The Windows ping and tracert command line utilities already have
options -4 and -6 to force one or the other protocol and I find it
very intuitive and convenient. By default they try IPv6 if the
target is/has an IPv6 address, and use IPv4 if the target does not
have an IPv6 address or if forced by option -4  (or /4 ).

A common use of a ping is just to very if a machine is somehow
reachable over the network. Often this suffices and it does not matter
over which PF it is reachable.

The ping6 and traceroute6 could be made just links to a common utility
and turn on the -6 option implicitly. For a transition period I could
live with the -4 being a default, if that would be a consensus.

I'm aware that ICMP and ICMP6 are quite different protocols, but from
a user's or sysadmin's perspective one should not need to check first
the local and remote protocol family of a host to be pinged, just to be
able to see if it is there. Tools like Nagios could benefit too.

I sincerely hope that these two utilities could be merged some day,
better sooner than later.

   Mark



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?a46d4e716d15f48b1d4bd32c22dce45c>