DHCP - Only leasing 39 of 150 addresses. Needs reboot.

[DCUK6]

Have a weird issue with DHCP.

 

Every so often it will say its run out of IP addresses. Looking in the leases its only using 39 of the 150 addresses in the scope.

 

I give the service a restart and it seems to dish them out again.

 

Havnt yet managed to get a log as to why its doing this.

 

There arnt any BAD_ADDRESS in the leases.

Can see in the log its slowly using them up.

 

Have set the lease time to 1 hour and 20 mins to try and get it to reuse.

 

Any thoughts?

[kmount]

Don't have any rogue exclusion ranges in place do you within the scope itself?

 

(i.e. the scope is 150 addresses but you've excluded 141 of them)

[mikkydoos]

Whats your range and the subnet mask value?

[DCUK6]

Sorry for the late replies, the non IT jobs keep coming (Long story).

 

We are on 255.255.254.0 with *.*.176-177.0

 

DHCP is on the *.*.176.2-150

 

Reservations are on the same scope but use the 177 range

 

Scope is named *.*.176.0 because it only used to be this range.

 

Has been working fine for 4-5 months but the last few weeks its starting to play up

 

 

 

 

*=Hidden IP

[Michael]

I agree, you need to check your scope, but also the distribution pool which could be set to anything.

 

I presume reverse DNS zones have been configured in DNS server?

[mikkydoos]

Also have you got DHCP configured on another server or DC also either in hot standby or load balancing?

 

I would try deleting all the leases and watch live what happens tomorrow morning.

[DCUK6]

Just going through active IP's now everyone has gone home and there about 7-8 i can set to out of the distribution.

Also the scope missed the last 10 IP's.

Weird though that when its saying it has 0 left i cant see all ip's in the leases. Only 1 device on the network that has a static IP in the range and that's our boiler. This does show as a bad address.

 

Reverse lookup isnt setup, something i can look into.

[Michael]

Another possibility of course is to delete and re-create the scope altogether, best out of hours also.

[Michael]

Generally speaking also, a restart of the DHCP Server service is normally suffice, rather than the whole server. I have numerous 2016 servers with the DHCP role working happily, so I wonder if there are other problems? Depending on your setup, (if using Hyper-V), should be easy enough to setup a new instance and point clients to look at it instead.