Date: Thu, 6 Oct 2016 15:52:53 +0200 From: "Hartmann, O." <ohartman@mail.zedat.fu-berlin.de> To: Oliver Peter <lists@peter.de.com> Cc: FreeBSD CURRENT <freebsd-current@freebsd.org>, allanjude@freebsd.org Subject: Re: CURRENT: "service netif restart" looses default route Message-ID: <20161006155253.3e97f90b@hermann> In-Reply-To: <20161006072753.GA11735@mail.opdns.de> References: <20161005184748.01aca03e.ohartman@zedat.fu-berlin.de> <20161006072753.GA11735@mail.opdns.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 6 Oct 2016 09:27:53 +0200
Oliver Peter <lists@peter.de.com> wrote:
> On Wed, Oct 05, 2016 at 06:47:48PM +0200, O. Hartmann wrote:
> >
> > Today, I checked on two servers of ours running both a recent
> > CURRENT (i.e. FreeBSD 12.0-CURRENT #43 r306701: Wed Oct 5 06:40:40
> > CEST 2016) via "service netif restart" the upcoming network and
> > realised that the default route is lost then!
> >
> > I'm able to config the route via "service routing restart" - or
> > manually as I did otherwise. But I recall that I did a simple
> > "service netif restart" in 11-CURRENT recently and that worked.
> >
> > Has there been a change? What is now the official way to restart
> > network?
>
> Since the past couple of years on every new FreeBSD I put this in
> motd for my linux colleagues and coworkers:
>
> Network:
> To apply changes you have made to the network:
> # /etc/rc.d/netif restart && /etc/rc.d/routing restart
>
> Perhaps we could introduce a wrapper to be used with:
> # service network restart
>
> oliver@local ~ % cat /etc/rc.d/network
> #!/bin/sh
>
> # PROVIDE: network
> # REQUIRE: netif routing
> # KEYWORD: nojailvnet
>
> . /etc/rc.subr
> . /etc/network.subr
>
> name="network"
> start_cmd="network_start"
> stop_cmd="network_stop"
>
> network_start()
> {
> /etc/rc.d/netif start
> /etc/rc.d/routing start
> }
>
> network_stop()
> {
> /etc/rc.d/netif stop
> /etc/rc.d/routing stop
> }
>
> load_rc_config $name
> run_rc_command $*
>
>
I'm sorry that my sloppyness brought up this discussion.
I recall that I tried to hunt down a bug and did lots of "service netif
restart". But on a local network with all essential servers on the same
net, I never faced a problem with routing.
If life gets too easy, people start digging for the reason of a problem
too deep when the framework does too much stuff automated ...
The knowledge of the fact that I also have to restart routing via a
script is sufficient for me - also the fact to know that netif restart
doesn't restart routing.
Regards,
Oliver
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20161006155253.3e97f90b>