On 10/7/21 9:52 pm, Geert Stappers via Dnsmasq-discuss wrote:
On Sat, Jul 10, 2021 at 05:57:40PM +1000, Paul Gear wrote:
On 10/7/21 5:13 pm, john doe wrote:
On 7/10/2021 7:58 AM, Paul Gear wrote:
Hi all,
It's been a while since I originally posted about this, and I've done a
lot of experimenting without landing on a working setup.
I'm trying to get a very simple test case working, where I have a
number
of vnet interfaces, all configured with the link-local address
169.254.0.1/16, and have dnsmasq allocate the leases out of a different
range (in this case, I'm using a portion of the documentation prefix,
203.0.113.0/29). Here's my configuration:
pid-file=/var/lib/dnsmasq-vnet/dnsmasq-vnet.pid
bind-dynamic
interface=vnet*
log-dhcp
no-hosts
dhcp-range=203.0.113.0, 255.255.255.248, 1d
From (1):
" this dhcp-range MUST include the netmask."
Unless I'm missing something, I don't see a subnet mask being specified.
1) https://thekelleys.org.uk/dnsmasq/docs/dnsmasq-man.html
Hi John,
Unless I'm missing something 255.255.255.248 is a netmask, and dnsmasq
doesn't appear to support CIDR notation in dhcp-range specifications. At
least, when I attempt to use /29 instead, it fails to start, saying:
dnsmasq: bad dhcp-range at line 41 of /var/lib/dnsmasq-vnet/dnsmasq.conf
Snippet dnsmasq manual page
--dhcp-range=[tag:<tag>[,tag:<tag>],][set:<tag>,]<start-addr>[,<end-addr>|<mode>][,<netmask>[,<broadcast>]][,<lease
time>]
So
dhcp-range=203.0.113.0, 255.255.255.248, 1d
should be something like
} } } dhcp-range=203.0.113.0, 203.0.113.7,255.255.255.248, 1d
Back to
From (1):
" this dhcp-range MUST include the netmask."
Unless I'm missing something, I don't see a subnet mask being specified.
1) https://thekelleys.org.uk/dnsmasq/docs/dnsmasq-man.html
If so, the
--dhcp-range=[tag:<tag>[,tag:<tag>],][set:<tag>,]<start-addr>[,<end-addr>|<mode>][,<netmask>[,<broadcast>]][,<lease
time>]
should be
--dhcp-range=[tag:<tag>[,tag:<tag>],][set:<tag>,]<start-addr>[,<end-addr>|<mode>],<netmask>[,<broadcast>][,<lease
time>]
or even
--dhcp-range=[tag:<tag>[,tag:<tag>],][set:<tag>,]<start-addr>,<end-addr>|<mode>,<netmask>[,<broadcast>][,<lease
time>]
Hi John & Geert,
So bottom line is that end-addr is not optional when netmask is
specified, even though a literal reading of the current man page
suggests that it is. I'll try to get a documentation patch together(is
here the right place to post it?) when I'm done.
That one little change was enough to turn my basic test case into a
fully working setup. Thanks for your help.
Regards,
Paul
_______________________________________________
Dnsmasq-discuss mailing list
Dnsmasq-discuss@lists.thekelleys.org.uk
https://lists.thekelleys.org.uk/cgi-bin/mailman/listinfo/dnsmasq-discuss
