Troubleshooting the DHCP Server

When problems occur on a DHCP server, it is worth checking whether the server itself is functioning before assuming the DHCP configuration is at fault. Use standard Solaris commands to verify that the server is working, such as ping , ifconfig , and snoop .

You can also stop and start the DHCP server process, in.dhcpd , as described in Chapter 11, "Basic DHCP." Doing this can often solve a number of problems, particularly when the process has hung and is not responding to client requests , or when you are using either the binary or nisplus format for the DHCP datastore, neither of which is human readable.

Error Messages

There are a number of reasons why a DHCP client might be unable to obtain an IP address from a DHCP server. Table 12.2 lists some of the more common errors encountered , along with a brief explanation of what the cause of the error is and also the required resolution. Note that the IP addresses will be different from those shown here.

Table 12.2. Common DHCP Error Messages

Message	Description	Resolution
No more IP addresses on 192.168.28.0 network	All the IP addresses that are managed by DHCP are in use for the specified network.	Use dhcpmgr or dhcpconfig to create additional IP addresses to allocate to clients .
Client:010800208E48CE is trying to renew 192.168.28.32, an IP address it has not leased	The IP address held by the server for this client does not match the IP address specified by the client. This is often caused by a client's record being deleted while the IP address is still in use.	Use dhcpmgr or pntadm to view the network table and correct.
Offer expired for client: 192.168.28.32	The DHCP made an offer of an IP address, but the client did not respond in time.	The cache offer timeout on the DHCP server may be set too low. Increase the value by using dhcpmgr .
192.168.28.32 currently marked as unuseable	The IP address has been marked in the network table as unuseable.	Use dhcpmgr or pntadm to make the address useable.
There is no 192.168.28.0 dhcp-network table for DHCP client's network	A client is requesting an IP address, but the DHCP server cannot find the network table for the address, usually because the network table has been deleted.	Re-create the network table by using dhcpmgr or dhcpconfig .
DHCP network record for 192.168.28.32 is unavailable, ignoring request	The specified IP address is not present in the network table.	Use dhcpmgr or pntadm to view the table and create the record if necessary.
ICMP ECHO reply to OFFER candidate: 192.168.28.32, disabling	The IP address is already in use. The server carries out this check before finally allocating the IP address to the client.	You need to establish whether the specified IP address should be managed by DHCP, or whether it is a static IP that is outside the scope of DHCP addresses.

Running the DHCP Server in Debug Mode

If there are problems with the DHCP server that cannot be easily resolved, it is useful to run the server process, in.dhcpd , in debug mode to receive additional messages on the server console. The following example shows the process running in debug mode, using the -d option, and also with verbose output, using the -v option. The lines of interest appear in bold.

 # /usr/lib/inet/in.dhcpd -d -v 3ef739e5:  Daemon Version: 3.5 3ef739e5:  Maximum relay hops: 4 3ef739e5:  Run mode is: DHCP Server Mode. 3ef739e5:  Datastore resource: SUNWfiles 3ef739e5:  Location: /var/dhcp 3ef739e5:  DHCP offer TTL: 10 3ef739e5:  ICMP validation timeout: 1000 milliseconds, Attempts: 1. 3ef739e5:  Name service update enabled, timeout: 15 seconds 3ef739e5:  Maximum concurrent clients: 1024 3ef739e5:  Maximum threads: 256 3ef739e5:  Read 3 entries from DHCP macro database on Mon Jun 16 18:33:25 2003 3ef739e5:  Monitor (0003/hme0) started... 3ef739e5:  Thread Id: 0003 - Monitoring Interface: hme0 ***** 3ef739e5:  MTU: 1500    Type: SOCKET 3ef739e5:  Broadcast: 192.168.28.255 3ef739e5:  Netmask: 255.255.255.0 3ef739e5:  Address: 192.168.28.28  3ef73a71:  Datagram received on network device: hme0   3ef73a71:  Reserved offer: 192.168.28.32   3ef73a72:  Unicasting datagram to 192.168.28.32 address.   3ef73a72:  Adding ARP entry: 192.168.28.32 == 0800208E48CE   3ef73a72:  Updated offer: 192.168.28.32   3ef73a73:  Datagram received on network device: hme0   3ef73a73:  Client: 010800208E48CE maps to IP: 192.168.28.32   3ef73a73:  Unicasting datagram to 192.168.28.32 address.   3ef73a73:  Adding ARP entry: 192.168.28.32 == 0800208E48CE  

The lines in bold show the messages received when a DHCP client requests an IP address. In this case, the IP address allocation worked because the client was mapped to the IP address successfully.

Later on, when the client is shut down, the debug output shows the same IP address being released:

  3ef73b2d:  Freeing offer: 192.168.28.32  3ef73b2d:  Datagram received on network device: hme0  3ef73b2d:  Client: 010800208E48CE RELEASED address: 192.168.28.32   3ef73b2d:  RELEASE: client message: DHCP agent is exiting