Thu, 23 Dec 2021 10:29:33 -0800 Andrea Bolognani <abologna(a)redhat.com>:
On Tue, Dec 21, 2021 at 12:22:44PM +0100, Olaf Hering wrote:
> --- a/src/remote/libvirtd.sysconf
> +++ /dev/null
> @@ -1,21 +0,0 @@
> -# If systemd socket activation is disabled, then the following
> -# can be used to listen on TCP/TLS sockets
> -#LIBVIRTD_ARGS="--listen"
This bit about passing --listen to libvirtd got lost during the move
to the unit file. Please make sure you preserve it.
It was wrong from day one to have this fragment here.
Documentation belongs to the doc directory.
> +++ b/tools/libvirt-guests.sh.in
> @@ -30,13 +30,53 @@ test ! -r "$sysconfdir"/rc.d/init.d/functions ||
>
> export TEXTDOMAIN="@PACKAGE@" TEXTDOMAINDIR="@localedir@"
>
> +# URIs to check for running guests
> +# example: URIS='default xen:///system vbox+tcp://host/system
lxc:///system'
> URIS="default"
[...]
This is the bit I still not entirely convinced of.
After this change, libvirt-guests' configuration becomes entirely
opaque: the only way the admin can learn how to configure the service
is by somehow realizing that it's a shell script as opposed to a
binary and looking inside it.
Can we do better?
libvirt-guests is undocumented. The file type is easy to guess, based on the filename in
libvirt-guests.service.
The Documentation= setting in this file is wrong, libvirtd(8) says nothing about this
functionality.
Initial documentation has to be provided as a separate change, by creating a new
docs/whatever.rst.
Olaf