Skip to content

Est. 2011ยทMicrosoft Partner ID 7033487ยทDelivery under 3 minยทSupport 7 days a week

Your vault is empty.

License Error Event ID 14

DHCP Event ID 14: the scope address pool is exhausted and leases now fail

10 min read Updated October 5, 2026 Windows Server: AD, DNS & Group Policy

Fix it now

ID 14 in the DHCP audit log is arithmetic, not a fault: every usable address in the scope is leased, reserved or excluded, so the next client is turned away. You either free addresses that are being held for devices that have gone, or you make the pool bigger.

Run these on the DHCP server in an elevated PowerShell session

Get-DhcpServerv4ScopeStatistics
Get-DhcpServerv4FreeIPAddress -ScopeId 10.0.0.0
Set-DhcpServerv4Scope -ScopeId 10.0.0.0 -LeaseDuration 08:00:00
Get-DhcpServerv4Lease -ScopeId 10.0.0.0 -AllLeases
  1. Compare the in-use count against the number of devices you believe are on site. Far more leases than devices is a lease-duration problem; a match is a subnet-size problem, and they have different fixes.
  2. Read the audit log for the day in %windir%\System32\Dhcp. ID 14 lines give you the times; ID 13 lines tell you addresses were taken out of circulation because something was already answering on them.
  3. Shortening the lease only helps as clients renew, because existing leases keep the duration they were issued with. Delete leases you know are dead with Remove-DhcpServerv4Lease if you need addresses back today.
  4. If the subnet has room, extend the range with Set-DhcpServerv4Scope -ScopeId 10.0.0.0 -EndRange 10.0.0.250 and re-check the free count.

Changing a subnet mask touches every static host, gateway, firewall rule and access list on that network. Plan it as a change window, not as a quick fix during an outage.

If clients are getting leases again you can stop here. The next section covers where the addresses went, which is rarely where people look first.

Why it happens

The DHCP audit log is a comma-delimited text file, one per day, written to %windir%\System32\Dhcp. It uses small numeric identifiers of its own, which are not Windows event IDs. ID 14 is the entry written when a lease request cannot be satisfied because the address pool of the scope was exhausted.

The usable size of a scope is the range, minus exclusions, minus reservations. Lease duration is the other half of the sum, and it is the half that goes wrong quietly. With an eight-day lease, a visitor who connects for ten minutes holds an address for eight days. On a guest or bring-your-own-device network the scope shows as full while the car park is empty.

Two other things shrink the pool without anybody deciding to. ID 13 records an address found in use on the network, which takes that address out of circulation – usually a statically addressed host sitting inside the dynamic range. And on the Windows event log side, Event ID 1063 says there are no addresses available for lease in the scope or superscope, while Event ID 20287 records a client request dropped because the applicable ranges are out of addresses, including ranges attached to a DHCP policy. A policy with its own narrow range can starve while the scope as a whole still has space.

The lease is far longer than devices stay

You have this one if The lease count is several times the number of devices on site, and most leases were last renewed days ago.

  1. Shorten the scope’s lease: Set-DhcpServerv4Scope -ScopeId <scope> -LeaseDuration 08:00:00 for an office network, shorter for a guest one.
  2. Expect gradual relief. Existing leases keep the duration they were issued with and only shrink as clients renew.
  3. Delete leases you know are dead with Remove-DhcpServerv4Lease if you need capacity today.

A short lease means more renewal traffic and a harder dependency on the server being reachable. Hours rather than minutes is the sensible range for most networks.

Addresses have been lost to bad address entries

You have this one if The free count is far below the range size, and the lease list contains entries in a bad-address state. ID 13 lines appear in the audit log.

  1. List everything with Get-DhcpServerv4Lease -ScopeId <scope> -AllLeases and read the address state on each entry.
  2. Find the statically addressed hosts answering inside the dynamic range, and either move them or exclude their addresses properly.
  3. Review conflict detection attempts on the server’s properties, Advanced tab: every attempt above zero costs a probe per lease and generates these entries on a noisy network.

A DHCP policy has run out of range, not the scope

You have this one if Some device types get leases and others do not, and the scope statistics still show free addresses.

  1. Look for Event ID 20287 in the DHCP-Server admin channel. It names the scope or superscope and says the applicable ranges are out of addresses.
  2. Open the policies on that scope and read the range set on each one, then widen it or remove the policy.
  3. Retry from a device that matches the policy condition.

The subnet is simply too small

You have this one if The lease count matches the real device count, and shortening the lease frees almost nothing.

  1. Count what the site actually needs: phones, printers, cameras and access points, not just computers.
  2. Re-subnet to a larger mask and update the scope, the router interfaces, the relay configuration and every static host.
  3. Where that cannot happen quickly, a superscope with a second range on the same physical network is a legitimate interim step – but only if the router carries an address in both ranges and the relay forwards for both.

Full reference

Reading the audit log without guessing

Audit ID What the line means
13 An IP address was found in use on the network
14 A lease request could not be satisfied because the address pool of the scope was exhausted
30 A DNS dynamic update request was sent
31 A DNS dynamic update failed
32 A DNS dynamic update succeeded

The files are named per day and kept in %windir%\System32\Dhcp. A day full of ID 14 lines with very few ID 10 lines (a new address leased) tells you the pool has been dry all day rather than momentarily tight at nine in the morning.

Windows event log entries that mean the same thing

Event ID Channel What it records
1063 DHCP-Server admin There are no IP addresses available for lease in the named scope or superscope
20287 DHCP-Server admin A client request was dropped because the applicable ranges in the scope or superscope are out of addresses, possibly a policy range
1002 DHCP-Server admin The DHCP service failed to initialise its global parameters. This is a service start failure and is unrelated to a full pool

Devices that should not be there

  • Give visitor and personal devices their own VLAN and scope with a short lease, so they cannot exhaust the network the business runs on.
  • Where a separate VLAN is not possible, separate device types with DHCP policies, each with its own range.
  • Look for a single adapter cycling through hardware addresses: it shows as a run of leases created seconds apart with names you do not recognise.
  • Re-check after a full working day rather than immediately – one afternoon is not a pattern.

Sizing the fix rather than repeating it

  • Work out peak concurrent devices, not headcount. A site of 100 people with laptops, phones and a printer each needs three times what a naive count suggests.
  • Add the lease overhang: with an eight-hour lease, devices that left in the last eight hours still hold addresses.
  • Leave exclusions for infrastructure at the bottom of the range and keep reservations out of the middle of it.
  • If the scope is failover-enabled, remember the standby server holds a reserve percentage of the pool, five per cent by default, so the usable pool on the active server is slightly smaller than the range suggests.

What not to do

  • Do not delete leases wholesale to free space. You will hand out addresses that are still in use and spend the afternoon on conflicts.
  • Do not set a lease of a few minutes on a general network. Renewal traffic rises and every DHCP hiccup becomes an outage.
  • Do not add a second scope for the same subnet on the same server; that is what a superscope is for, and it still needs routing support.
  • Do not resize the range up to the broadcast address. Leave the gateway and any reserved infrastructure addresses outside the pool.

When a licence is the actual fix

Nothing in the paragraphs above needs a purchase: resizing a scope, shortening a lease and clearing bad address entries are free, and they are the fix. Licensing only enters when the pool ran dry because the estate genuinely grew, because Windows Server is licensed per server with a client access licence for each user or device that uses it. Device CALs are the cheaper shape where many people share fewer machines – shift work, shop floors, wards, classrooms – and user CALs suit the opposite pattern, one person with a laptop, a phone and a desk PC. Arco supplies Windows Server Device CALs and will work through which of the two counts fits how your users and devices actually pair up, rather than selling the larger number.

Every code this article covers

Code What it points at Source
Event ID 14 DHCP audit log: a lease request could not be satisfied because the address pool of the scope was exhausted Microsoft Learn
Event ID 13 DHCP audit log: an IP address was found in use on the network, so it is taken out of circulation Microsoft Learn
Event ID 1002 The DHCP service failed to initialise its global parameters. A service start failure, not an exhaustion event – the exhaustion events are 1063 and 20287 Microsoft Learn

Confirm the fix worked

  1. Get-DhcpServerv4FreeIPAddress -ScopeId <scope> returns addresses.
  2. Get-DhcpServerv4ScopeStatistics shows the percentage in use dropping and holding.
  3. A test client releases and renews and gets a lease immediately.
  4. The next day’s audit log contains no new ID 14 lines.
  5. No new 1063 or 20287 events appear in the DHCP-Server admin channel.

Questions people ask about this

Does an exhausted scope mean I have to buy something?

No. Making the scope bigger or the lease shorter costs nothing. The only licensing consequence is indirect: if you now have more devices using Windows Server than you have client access licences for, that gap needs closing regardless of DHCP.

How short can a lease safely be?

Short enough that addresses come back before the next intake, long enough that the network survives a DHCP restart. Hours suit guest networks, a day or two suits offices, and minutes create renewal traffic for no benefit.

Will a superscope solve this?

Only if the routing supports it. A superscope lets one physical network carry two ranges, but the router interface needs an address in both and the relay has to forward for both. Without that, clients get an address they cannot use.

Why does the scope show as full when half our staff are away?

A lease is held for its full duration whether the device is present or not. Shortening the duration is what converts absent devices back into free addresses.

Some devices get addresses and others do not. Is the scope really full?

Check for Event ID 20287 and look at your DHCP policies. A policy carries its own range, and that range can be exhausted while the scope still has space.

Related error codes

Was this article helpful?

Your feedback helps us improve our documentation.

Related articles

Free Fix Error 0x80094012: certificate template permissions deny the enrolling account License Error Error 8557 machine account quota exceeded: no more computers may join License Error Error 8568: the functional level will not rise while legacy DCs remain Free Fix Event ID 1058: Windows cannot access gpt.ini for a Group Policy object
โ† Back to Knowledge Base