DHCPD-7-NO_LEASE: DHCP assignment failure, NO POOL

[TheGoob]

Hello

Not sure what is causing this and really not sure how to troubleshoot. Router was rebooted about 3 days ago and never had issues until today.

Is DHCPD-7 referring to the DHCP Pool #7 [I do have] or is that the standard error # and also it mentions client 000c.2968.65cb reason NO POOL.

8000v Router

[MHM Cisco World]

No. 7 is refer to no. Of log level.

MHM

[TheGoob]

Makes sense.. With that, does the error refer to the lease time? I never modified that... Would that be something to change?

[TheGoob]

Having a time here trying to isolate this.

Here is anything pertaining to DHCP;

ip dhcp excluded-address 10.0.1.1
!
ip dhcp pool LAN
 network 10.0.1.0 255.255.255.0
 domain-name esxi.router
 default-router 10.0.1.1
 dns-server 1.1.1.1
 lease 7
!
interface GigabitEthernet4
 ip address 10.0.1.1 255.255.255.0
 ip nat inside
 negotiation auto

[Devaa]

Could you please share "show ip int br | ex unass" and check how the DHCP Discover from that client hits the Router. Ideally it should hit the router via interface GigabitEthernet4

[TheGoob]

show ip int br | ex unass
Interface IP-Address OK? Method Status Protocol
GigabitEthernet2 10.0.0.1 YES NVRAM up up
GigabitEthernet3 192.168.4.2 YES DHCP up up
GigabitEthernet4 10.0.1.1 YES NVRAM up up
Dialer1 207.108.121.182 YES IPCP up up

[Jens Albrecht]

As expected you have a few more active interfaces so that it is clear that you got a DCHP request on one of those.

There is no need to explicitly have a DHCP client connected, simply booting an unconfigured device is enough to get a request due to the default autoconfiguration feature.

[TheGoob]

Well still confused;

GE2 is just the link between Router and SG500X Switch, and on the Switch I have vlans and DHCP Servers there, none of which are associated with the 10.0.0.0/24 Network, as it is simply a L3 to link to SG.

GE3 is the "management" Interface which obtains an IP from esxi Switch, again not sure why DHCP on that would interfere with the router, as it grabs it's IP from the higher level esxi switch, or would it also try to grab one from the switch, causing this issue...So should I set GE3 to static 192.168.4.2

GE4 is just an interface going to it's own L2 switch and anything connected to that switch would grab a 10.0.1.0/24 IP via DHCP Server on the Router.

Or am I missing the bigger picture?

[Jens Albrecht]

Well, it does not matter whether any other link is associated with the 10.0.1.0/24 network.
It also does not matter whether you configured static IP addresses on the other interfaces of the Cat8000v or use DCHP.

So let's take a look at the sample topology that I used to create the debug output that I posted previously:

The Cat8000v router is configured as DHCP server for the 10.0.1.0/24 network, serving IP addresses to the clients on the right side.

The Cat8000v router has a static IP address in the 172.16.1.0/24 network, while the Server on the left is doing DHCP for the client on the left side.

As soon as ubuntu-2 issues a DHCP request, I get this "NO POOL" message on the Cat8000v router.

Why? Because the Cat8000v router listens for DHCP requests on ALL of its active interfaces, once you configured the pool for interface GE1.

So what does this mean for your topology...
You mentioned that Cat8000v is using DHCP to get its IP address on GE3, right?
It is safe to assume that the DHCP server for this subnet does not only serve Cat8000v but other clients on this subnet. So as soon as any other client in this subnet issues a DHCP request your Cat8000v will receive it as well and issues the "NO POOL" message.
As mentioned above, changing GE3 to a static IP address does not change anything, as requests from other clients will still reach Cat8000v's interface GE3 causing that message.

Did you check the "debug ip dhcp server events" command on your Cat8000v?
Once you get the "NO POOL" message, it will tell you

• the incoming interface
• the subnet for which the DHCP request was issued
• the MAC address of the DHCP client

HTH!

[TheGoob]

Well that indeed makes a lot more sense... I did not realize that it would scan for DHCP requests, or get them, from interfaces not part of the initial DHCP Server, but I can see how it would.

Well oddly enough before we got into more in depth conversation, I have rebooted the router, the switch and my internet and since then no notices..So apparently I cleared out whatever issue it was but the debug is indeed on so if it comes up again I will be able to let you know the details. I thank you very much

[TheGoob]

Oddly enough I have not had that notice since I rebooted. But I do notice every 2 minutes is 'DHCPD: checking for expired leases'. Is this the correct duration for it to check?

[Jens Albrecht]

Yepp, that's the default behavior.

Depending on the lease time you configured for your pool you can adjust the timer using the command "ip dhcp binding cleanup interval <Seconds>". The default timer is based on the default lease time of 1 day. If you changed the lease time to e.g. 3 or 4 weeks, then you can also adjust the cleanup timer to let's say 600 seconds.

Background:
When a lease expires, it is not automatically removed from the DHCP database right after the expiration time.
Only during the next cleanup cycle, the DHCP process will remove any expired bindings.
The timer for this cleanup cycle is configured as mentioned above.

BTW, you mentioned the MAC address of the client causing the "NO POOL" message last time.
So you should be able to locate this client e.g. by checking the MAC address-table of your switches or your inventory list if you created one.

HTH!

[Jens Albrecht]

The important key word of your error message is "NO POOL".

Your DHCP pool is created to serve requests coming in on interface Gi4.
So if your router receives a dhcp request on any other interface, e.g. Gi1 - 3, then you get such error message. It means that the router receives a dhcp request on a subnet that does not have a dhcp pool linked to it.

The router also tells you the client mac address so that you can check the mac address-table of your connected switches to find the location of the client and where it is connected.

You can also use the command "debug ip dhcp server events" to get more details, telling you the incoming interface and subnet:

Hope this helps!