DHCP not working on non native vlan

Hi

I am having a pretty fundamental problem with my Route10. The default vlan 1 is 192.168.0.0/24 with gateway .1. I can get wired and wireless clients attached to this vlan. Clients directly attached to the Route10 and my 8 port switch get an address and can get to the internet.

I have another vlan I’ve added to the Route 10, 192.168.254.0/24 vlan 254. I’ve created a new SSID and attached it to this vlan but I cannot get a DHCP address. Even if I assign a static IP I cannot ping my gateway or the internet.

I’m very short on cli troubleshooting commands but when I did

arp -a

I can see a wired client I have on vlan 254 with a static IP. The arp entry for the IP is correct and the IP is in the correct vlan.

The link from my Route10 to my switch is default native vlan 1 and vlan 254 is allowed on the link at both ends. So vlan 254 should be tagged on the interlink and untagged on a switch port connected to my wired client. Still no DHCP address.

All the settings look pretty standard. DHCP is enabled on vlan 254, I have left the default 10 IP addresses reserved and added a specific DNS server address.

Anyone have any ideas ? Either I’m missing something pretty fundamental or the Route10 is not working properly.

JS.

Here’s my additional subnet settings

Here’s how I attach the vlan to my SSID

Update.

Strangely I created another subnet, vlan 253 - 192.168.253.0.24, and attached it to my SSID and I immediately connect and get a DHCP address. I then deleted my vlan 254 and re-added it but it still refuses to dish out an IP address. I added a further vlan 252 - 192.168.252.0/24 and it works. The route 10 obviously doesn’t like vlan 254 or subnet 192.168.254.0/24

Seems like some sort of reserved, fallback, managment or similar address/subnet, even though there is no clear documentation to support that.

Did you try 10.10.254.0/24 and .1/24 respectively.

Works perfectly fine with 10.10.254.1/24 as my vlan 10 or vlan 254 address. In fact it gets stranger. I have deleted 192.168.254.1/24 and recreated it. Now I do get an IP address in the range and can get to the internet but I cannot ping my gateway 192.168.254.1. A traceroute gets no response from my gateway but subsequent hops are ok.

If I change my subnet to vlan 253 - 192.168.253.1/24 I can ping my gateway.

Also, I notice that by default 10 IP addresses are reserved from the pool but the first device I put on the vlan always gets .10 as an address. I would have expected .11 to be the first available dhcp address.

The counting begins at 0, so ends at 9.

Is it really appropriate to have a DHCP-range starting at 0, when setting Reserved IPs=0, or starting at 1 when setting 1? Or is that common practice?

Edit: It seems awkward to have to configure it as for example 5 to get exactly 3 (=5-2) available reserved IPs.

I was trying to figure out a way to reply to this… you can’t assign a device .0 so why on earth would you start the count there??

2 Likes

It doesn’t really start at 0. It starts at the first address in the subnet, which is usually zero for a typical network. However, that subnet can still be subdivided further such as 192.168.254.1/25 + 192.168.254.129/25.

Capturing the nuance and necessity of counting from 0 in the tooltip is difficult. Perhaps a feature request to preview the reserved/dhcp ranges would be valuable?

Yes that as a “feature” is extremely important. It shouldn’t even need to be a feature request, it should be that way from the beginning. It’s already confusing enough.

But, it must surely be a bug to allow .0 and .1 in on the DHCP range?!

.0 will never actually be used, even if it shows in the UI. The DHCP server will ensure that. .1 is valid if you use .254 as a router, etc.

I’m trying to understand the desire of setting Reserved IPs to 0 in the first place. We may just make 1 the minimum.

1 Like

As not being schooled in network technology, my first assumption was that reserved meant reserved for static IP clients, exclusive of the gateway IP, and contrary to dynamic IPs.

In other words, if I wanted all clients to have dynamic IPs allocated by DHCP, I would readily assume that reserved should be set to 0. Then, it gets even more confusing when the DHCP range is presented in the GUI as starting at .0.

So, a first measure would be to fix the GUI show the actual DHCP range, which would be at least one above the gateway IP.

Agree, perhaps output a red text warning for anything other than empty/blank or numeric at 1 or above.

Some additional text for the tooltip on the gateway being included in the reserved IPs, and that it must be one or higher, and that reserved IPs in excess to one would be the number of allowed static IPs, would benefit the understanding as well.

1 Like

One should be able to catch that if you happened to input 0 under reserved and save, it would clearly shows that network DHCP range starting at xxx.xxx.xxx.0