Australian cloud for business software — built and run in Sydney since 2002

NetLeverage
NetLeverage

DHCP scope full: clients get 169.254 addresses. Find what took the leases

Technical article
Troubleshooting

About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.

A new laptop, a visitor's phone or a replaced printer gets no network. ipconfig /all shows an address starting 169.254. Devices that were already on the network keep working, until their lease runs out.

A 169.254 address is one a Windows computer gives itself when it cannot get an address from a DHCP server. Microsoft describes this as Automatic Private IP Addressing: the range 169.254.0.0 to 169.254.255.255 is reserved for it, and a computer using it can talk only to others on the same local network that are also in that range (Microsoft Learn, Automatic TCP/IP addressing without a DHCP server). A scope with no free addresses left is one way a client ends up there; the checks below confirm whether that is the cause.

What you see

  • New or returning devices get 169.254 addresses; devices that already hold a lease are fine.
  • The scope statistics in the DHCP console show the scope at or near 100% in use.
  • The Address Leases list contains BAD_ADDRESS entries.
  • With two DHCP servers sharing the scope (failover), one server has addresses left and the other does not.

Why it happens

Microsoft's DHCP server checklist (Microsoft Learn, DHCP troubleshooting guidance) and its scope and failover pages point to most of these causes:

  • The range is too small for the devices on the subnet. Each subnet has one scope with one continuous range of addresses (Microsoft Learn, DHCP scopes). Exclusions and reservations come out of that range.
  • Declined addresses (BAD_ADDRESS). When a client declines an address because another device is already using it, the server marks that lease as bad. Microsoft's reference says such an address is not offered again until a 10-minute timer expires (Microsoft Learn, Get-DhcpServerv4Lease).
  • Static addresses that are not excluded. Microsoft's guidance is to exclude every address set by hand on servers, printers and other devices. If they are not excluded, the server and the static device can both claim the same address.
  • Leases that last longer than the devices stay. A scope's lease duration is set per scope. On a subnet where many devices come and go, addresses held by devices that have left are not free until their leases expire.
  • Devices that change their network (MAC) address. Apple says its phones, tablets and computers identify themselves to each Wi-Fi network with a private address and "might rotate (change) the address periodically" (Apple Support, Use private Wi-Fi addresses on Apple devices). A lease is normally tied to the client's hardware address. It follows that a device that changes it is given a new lease while the old one is still held until it expires, so on a busy guest or staff Wi-Fi scope, one device can hold several addresses at once.
  • Failover splitting the pool. In hot standby mode the standby server keeps a reserve of the addresses, 5% by default. If the active server is unreachable and the standby uses up that reserve before the maximum client lead time (MCLT) expires, it refuses new leases and only renews existing ones. In load balance mode the default split between the two servers is 50:50 (Microsoft Learn, DHCP failover).

How to prove the cause

1. Check the scope's usage. In the DHCP console, look at the statistics for the scope serving the client's subnet. In PowerShell (Microsoft Learn, Get-DhcpServerv4ScopeStatistics):

Get-DhcpServerv4ScopeStatistics -ComputerName <DHCP server> -ScopeId <scope ID>

Microsoft's own example lists scopes more than 80% used:

Get-DhcpServerv4ScopeStatistics -ComputerName <DHCP server> | Where-Object -FilterScript { $_.PercentageInUse -Gt 80 }

Add -Failover to include the failover statistics when the scope is shared with a partner.

2. Look for declined leases:

Get-DhcpServerv4Lease -ComputerName <DHCP server> -ScopeId <scope ID> -BadLeases

Each one is an address a client found already in use by another device, for example one set by hand inside the range.

3. List every lease, not just active ones, to see what is holding the range:

Get-DhcpServerv4Lease -ComputerName <DHCP server> -ScopeId <scope ID> -AllLeases

Microsoft notes the command returns only active leases unless -AllLeases is given.

4. Compare with your list of static devices. Any static address inside the scope range that is not in an exclusion range is a conflict waiting to happen.

5. Check the event logs and the DHCP server log. Microsoft points to the System log, the DHCP Server channels under Applications and Services Logs > Microsoft > Windows > DHCP-Server, and the server's debug logs in %windir%\System32\Dhcp.

6. Rule out other causes if the scope is not full: the server is authorised and running, the relay agent (IP helper) on the router is set for the client's subnet, and nothing else is using UDP ports 67 and 68 on the server.

The fix

  • Exclude static addresses. Add an exclusion range covering every device set by hand, as Microsoft's scope guidance recommends, leaving spare room for growth. Microsoft's example from its quickstart:

    Add-DhcpServerv4ExclusionRange -ScopeId 192.168.15.0 -StartRange 192.168.15.1 -EndRange 192.168.15.10
    

    Then move the static devices into that excluded block, or give them DHCP reservations. Live change: re-addressing a server, printer or other static device breaks anything that reaches it by its old address (shortcuts, print queues, firewall rules, DNS records set by hand) until those are updated. Do it in a planned window, one device at a time.

  • Shorten the lease on a subnet with many short-stay devices, such as guest Wi-Fi (Microsoft Learn, Set-DhcpServerv4Scope). Microsoft's example sets a four-day lease:

    Set-DhcpServerv4Scope -ComputerName <DHCP server> -ScopeId <scope ID> -LeaseDuration 4.00:00:00
    
  • Widen the range if the subnet has room. Microsoft's reference notes that setting a new start or end address discards the previous range, so give both. If the subnet itself is too small, the fix is a larger subnet or a new one, which means router and VLAN changes as well.

  • Turn on conflict detection if static devices keep causing declines. The server can test an address before leasing it; Microsoft's reference gives 0 to 5 attempts, default 0 (off) (Microsoft Learn, Set-DhcpServerSetting):

    Set-DhcpServerSetting -ComputerName <DHCP server> -ConflictDetectionAttempts 1
    
  • Fix the failover side. If a hot standby server ran out of its reserve because the active server was down, bring the partner back or follow your failover recovery steps. Microsoft notes that changes to a failover scope must be replicated to the partner by hand.

How to confirm the fix

  • Get-DhcpServerv4ScopeStatistics shows free addresses in the scope.
  • Get-DhcpServerv4Lease -BadLeases stops gaining new entries.
  • A device that had 169.254 gets a proper address after ipconfig /release then ipconfig /renew.

What to watch afterwards

  • Scope usage. Microsoft's own example filters scopes above 80% in use; an alert around that level leaves time to add addresses before the scope is full.
  • New static devices added inside the range without an exclusion.
  • Seasonal growth in phones, visitors and IoT devices on shared subnets.
  • Failover health, because a partner outage changes how many addresses each server can hand out.

Risks and limits

  • Shortening leases increases renewal traffic and makes clients more dependent on the DHCP server being up; change it on the busy subnet only.
  • Changing a scope's range is a live change. Note the current start, end and exclusions before editing.
  • Conflict detection makes the server test each address before leasing it, so each new lease takes a little longer.
  • Microsoft's automatic-addressing page was written for older Windows versions; it is used here only for what a 169.254 address means.
  • An authorisation fault or a broken relay agent can also leave clients on 169.254. Check those if the scope is not full.

Sources

Next step

A full scope is a capacity problem that shows up as a network fault. NetLeverage has looked after business servers since 2002. If devices on your network are getting 169.254 addresses, talk to us through managed support with the scope statistics and the bad-lease list from your DHCP server.