13    Solving Network and Network Services Problems

This chapter contains a diagnostic map to help you solve problems that might occur when you use the network and network services software. Use this chapter together with the appropriate Digital documentation to solve as many problems as possible at your level.

13.1    Using the Diagnostic Map

Network and network service problems can occur for a number of reasons. The diagnostic map in this chapter should help you isolate the problem. The following figure explains how to use the diagnostic map:

After you isolate the problem, the map refers you to other chapters for instructions on using the various problem solving tools and utilities. The map also refers you to other manuals for more complete diagnostic information for particular devices and software products.

You could experience problems that are not documented in this manual when you use base system network and network services software with other layered products. Each layered product has its own manual or set of manuals, which might describe other solutions not provided here.

13.2    Getting Started

Before you start problem solving, ensure that the communications hardware is ready for use. Verify the following:

• The system's physical cable connections (the Ethernet connection and the transceiver connection) are properly installed. See the documentation for your system and communications hardware device.

• Event logging is enabled in order to monitor network events. See System Administration for information on starting event logging and for descriptions of the event messages.

Also check the product release notes for up-to-date information on known problems.

Table 13-1 helps you identify a starting point in the diagnostic map.

Table 13-1: Problem Solving Starting Points

13.3    Solving Network Problems

	Turn on the power to your system. See the system manual for your system's startup procedure and any problem solving information.
	If you are running NIS and your system hangs after the NIS daemons are started and before it mounts remote file systems, no NIS server is available to respond to the ypbind request. If you know there is an NIS server for your domain, wait until the server responds; the boot procedure will continue. If there is a LAT problem, the following message appears: getty: cannot open "/dev/ttyxx" See the solutions for solving LAT problems in Section 13.14. If your system is an NFS client and it hangs while mounting a remote file system or directory, complete the following steps: Check the cable and connection between your system and the network. Wait until all the servers listed in the /etc/fstab file are available on the network; your system will then continue booting. If you want your system to continue booting even if an NFS server is down, do the following: Halt the system. Boot the system to single-user mode and run the fsck command on the local file systems. Edit the /etc/fstab file and add the bg (background) option to the server entries. See Chapter 8 for the correct format of an fstab entry with the bg option. Reboot the system with the following command: #  /sbin/reboot If the bg option is specified in the fstab file entry, the remote file system or directory is automatically mounted when the server is running and begins functioning as an NFS server.
	Check the /etc/rc.config file for the following entry: NUM_NETCONFIG If the entry is 0, run the Network Configuration application. See Section 2.3 for more information.
	Verify that the network daemon (inetd) is running. Enter the following command: #  ps -e | grep inetd If no inetd daemon is running, start it, using the following command: #  /sbin/init.d/inetd start
	If a remote host's network is not reachable, the following message is displayed: network is unreachable Complete the following steps: Ensure that the network devices are configured properly on the local host, using the netstat -i command. See Section 2.3 for information on configuring network devices. Check the routing tables on the local host, using the netstat -r command. Trace the path looking at each Internet Protocol (IP) router's routing tables to make sure there is an entry for the remote host's network. Repair the incorrect IP router's routing tables. (This step requires a thorough knowledge of your topology.) Check that the local host's address-to-name translation for the remote host is correct. See the solutions for [Host known?]. Check the routers along the path to the remote host to determine whether they have security features enabled that prevent you from reaching the remote host.
	If a remote host is not known, the following message appears: unknown host Complete the following steps: Check if the user is trying to reach the remote host using a valid host name. Check if the remote host is in another name domain and that the user specified the full domain name. If your site uses the BIND name service for name-to-address translation, check the /etc/svc.conf file to see if bind is specified as a service for the hosts database entry. If it is not, edit the file and add it. Also, verify if the BIND service has information about the remote host. See the solutions for solving BIND client problems in Section 13.5. If your site uses NIS name service for name-to-address translation, check the /etc/svc.conf file to see if nis is specified as a service for the hosts database entry. If it is not, edit the file and add it. Also, verify if the NIS service has information about the remote host. See the solutions for solving NIS client problems in Section 13.7. If your /etc/svc.conf file lists local as the only name-to-address translation mechanism, the /etc/hosts file does not have information on the remote host. See System Administration for more information.
	If a remote host is not reachable, the following message appears: host is unreachable Complete the following steps: Check the cabling between the local host and the network. Verify that the remote host is running, using the ping command. Make sure that the network devices are configured properly on the local host, using the netstat -i command. See Section 2.3 for information on configuring network devices. Check the routing tables on the local host using the netstat -r command. Use the ping command to determine whether the IP router is reachable. Check that the local host's address-to-name translation for the remote host is correct. See the solutions for [Host known?]. Check the routers along the path to the remote host to determine whether they have security features enabled that prevent you from reaching the remote host.
	If a file cannot be accessed using the rcp or rsh commands, the following message appears: permission denied Complete the following steps: Check that the user is intended to have access to the remote host. The remote host might be intentionally preventing remote access. Check that the correct host and user definitions exist in the user's .rhosts file on the remote host. Check that the /etc/hosts.equiv file has been set up correctly. Check that the directory and file protection on the files to be copied or the .rhosts file on the remote system are correct. If you are using NFS, go to Section 13.9.
Problem still exists? Report your problem to Digital. See Chapter 16.	If the connection is broken, the following message appears: connection timed out Complete the following steps: Test the network to determine whether the problem is on the local host, remote host, or a host on the path between the two. See Chapter 14 for more information on testing the network. Once you have identified the host with the problem, do the following: Verify that the broadcast address and address mask for the local host are properly set up in the /etc/rc.config file. Confirm that the network device is properly configured. Make sure the local host's /etc/hosts file has the correct IP address for the local host. Make sure the cabling from the local host to the network is intact and properly connected. If connected over a local area network (LAN), check the Address Resolution Protocol (ARP) entries and LAN connections. If connected over a wide area network (WAN), check WAN connections and modems.

13.4    Solving BIND Server Problems

	Verify whether the Additional Networking Services subset is installed. Enter the following command: #  setld -i | grep OSFINET The following message should appear: OSFINETnnn installed Additional Networking Services (Network-Server/Communications) If the subset is not installed, install it by using the setld command. See System Administration for more information on installing the subset.
	Check the /etc/rc.config file for the following entry: BIND_SERVERTYPE= If no type is specified (double quotation marks), run the BIND Configuration application. See Section 6.3 for more information.
	Verify that the BIND daemon (named) is running. Enter the following command: #  ps -e | grep named If no named process is running, start the named daemon, using the following command: #  /sbin/init.d/named start
Problem still exists? Report your problem to Digital. See Chapter 16.	If the nslookup command does not return information for any host or the host specified in the client nslookup command, check the /etc/rc.config file for the following entry: BIND_SERVERTYPE= If the type is: Go to: CLIENT Stop. This system is not a BIND server and cannot provide name resolution to clients. PRIMARY Section 15.3 SECONDARY Section 15.3 SLAVE Section 15.4 CACHE Section 15.8

13.5    Solving BIND Client Problems

	Verify whether the Additional Networking subset is installed. Enter the following command: #  setld -i | grep OSFINET The following message should appear: OSFINETnnn installed  Additional Networking Services (Network-Server/Communications) If the subset is not installed, install it by using the setld command. See System Administration for more information on installing the subset.
	Check the /etc/rc.config file for the following entry: BIND_CONF="YES" If the entry does not exist, run the BIND Configuration application. See Section 6.3 for more information.
Problem still exists? Report your problem to Digital. See Chapter 16.	If you attempted to use one of the network commands (for example, telnet, rlogin, and rsh commands) and the remote host is not known, the following message appears: unknown host Complete the following steps: Check the /etc/svc.conf file to determine whether BIND is being used for the hosts database lookup. If it is, go to step 2. If it is not, add it to the file by using the /etc/svcsetup command. Retrieve information about the remote host with which you tried to communicate by using the nslookup command. Enter the following command: #  nslookup hostname If the command succeeds, the client is set up correctly; try the network command again. If the command fails, go to step 3. View the /etc/resolv.conf file and retrieve the addresses for the nameserver entries. Verify whether the servers are reachable by using the ping command. If no servers are reachable, contact your network administrator. If any name server fails to respond to the ping command, delete the name server entry from the resolv.conf file. Try the nslookup command again. If the command fails, see the solutions for solving BIND server problems in Section 13.4.

13.6    Solving NIS Server Problems

	Verify whether the Additional Networking Services subset is installed. Enter the following command: #  setld -i | grep OSFINET The following should appear: OSFINETnnn installed  Additional Networking Services (Network-Server/Communications) If the subset is not installed or is corrupt, install it by using the setld command. See System Administration for more information on installing the subset.
	Check the /etc/rc.config file for the following entry: NIS_CONF="YES" If the entry does not exist, run the nissetup script. See Section 7.3 for more information.
	Verify that the portmap daemon is running. Enter the following command: #  ps -e | grep portmap If you do not find the portmap daemon, stop and restart NIS, using the following commands: #  /sbin/init.d/nis stop #  /sbin/init.d/nis start If the portmap daemon does not start, reboot the server.
	To verify that a ypserv process is running, enter the following command: #  ps -e | grep yp If no ypserv process is running, stop and start NIS, using the following commands: #  /sbin/init.d/nis stop #  /sbin/init.d/nis start If a ypserv process is running, execute a ypwhich command. Enter the following command: #  ypwhich If the ypwhich command does not return an answer, find the process ID (PID) of the portmap process and kill it. Enter the following commands: #  ps -e | grep portmap #  kill -9 portmap_PID Note Since other network services use the portmap daemon, stopping it can affect network service. Therefore, notify your users of potential disruptions. Stop and start NIS by using the following commands: #  /sbin/init.d/nis stop #  /sbin/init.d/nis start
	Verify the information in the map. Enter the following command: #  ypcat map_name Where map_name is the name of the NIS map. If the information is incorrect, create a new map. Enter the following commands: #  cd /var/yp #  make map_name The make command returns the following message: map_name updated If the make command indicates that the database is not updated, complete the following steps: Remove the database_name.time file in the /var/yp and /var/yp/domainname directories. Create a new map by using the make command. Enter the following commands: #  cd /var/yp #  make map_name
Problem still exists? Report your problem to Digital. See Chapter 16.	If you suspect that a slave server is not getting NIS map updates, complete the following steps on the slave server: Verify that the NIS master server is running and reachable, using the ping command. See Section 14.1 for more information on using the ping command. Create a ypxfr log file. Enter the following commands: #  cd /var/yp #  touch ypxfr.log Run ypxfr interactively to get map updates. Enter the following command: #  ypxfr mapname Check the ypxfr.log file. After checking the log file and resolving any problems, remove the log file. This turns logging off. Verify the ypxfr entries in the /var/spool/cron/crontabs/root file. Use either the pg command or the /usr/bin/crontab -l command. The slave server entries are similar to the following: # Network Information Service: SLAVE server entries 30 * * * * sh /var/yp/ypxfr_1perhour 31 1,13 * * * sh /var/yp/ypxfr_2perday 32 1 * * * sh /var/yp/ypxfr_2perday Verify that the map has an entry in the corresponding ypxfr shell scripts. Check the syslogd daemon message files for any NIS messages. See Section 14.8 for more information. Verify that the slave server is in the ypservers map for the domain.

13.7    Solving NIS Client Problems

	Check the /etc/rc.config file for the following entry: NIS_CONF="YES" If the entry does not exist, run the nissetup script. See Section 7.3 for more information.
	Check the svc.conf file to be sure that it has entries for NIS. Use the /usr/sbin/svcsetup script to verify this. NIS entries are indicated by the letters "yp." For the passwd and group databases, the Security Integration Architecture (SIA) controls whether NIS is used. However, in order to use NIS, the following characters must appear as the last line in both databases: +:
	Verify that the portmap daemon is running. Enter the following command: #  ps -e | grep portmap If no portmap daemon is running, stop and restart NIS, using the following commands: #  /sbin/init.d/nis stop #  /sbin/init.d/nis start If the portmap daemon does not start, reboot the client.
	Verify that a ypbind process is running. Enter the following command: #  ps -e | grep yp If no ypbind process is running, stop and start NIS, using the following commands: #  /sbin/init.d/nis stop #  /sbin/init.d/nis start If a ypbind process is running, execute a ypwhich command. Enter the following command: #  ypwhich If the ypwhich command does not return an answer, kill the portmap process. Enter the following command: #  kill -9 portmap_PID Stop and start NIS, using the following commands: #  /sbin/init.d/nis stop #  /sbin/init.d/nis start
	If the ypwhich command gives you inconsistent information when invoked several times in succession, your client system is changing the server system to which it is bound. This can occur over time, especially if your system is on a busy network or if the NIS servers are busy. Once all clients get acceptable response time from the NIS servers, the system will stabilize. If the ypwhich command reports that the domain is not bound, your system did not initially bind to a server system. Issue a ypcat command, then reissue the ypwhich command again.
Problem still exists? Report your problem to Digital. See Chapter 16.	If an NIS command hangs, the following message appears on the console: yp: server not responding for domain domainname. Still trying The client cannot communicate with the server. Complete the following steps: Verify that the domain name returned by the domainname command matches the NIS_DOMAIN variable in the server's /etc/rc.config file. If the domain name does not match, reconfigure the client system by using the nissetup script. See Section 7.3 for information on configuring NIS. Verify that there is at least one NIS server for your domain running on your local subnetwork. If there is not, you must use the -S option to the ypbind command. Reconfigure the client, using the nissetup command, and choose this option. Check with other clients on the subnetwork to determine if they are having problems with any of the NIS commands. Verify that the server daemons on the remote system are running. Enter the following command: #  rpcinfo server_name Check the syslogd daemon message files for any NIS messages. See Section 14.8 for more information. Verify that the server is running. See the solutions for solving NIS server problems in Section 13.6. If the previous steps do not solve the problem, complete the following steps: Stop and start NIS. Enter the following commands: #  /sbin/init.d/nis stop #  /sbin/init.d/nis start If this does not solve the problem, go to step 2. Reboot the system. Reconfigure NIS by running the nissetup script.

13.8    Solving NFS Server Problems

	Verify whether the NFS subset is installed. Enter the following command: #  setld -i | grep OSFNFS The following message should appear: OSFNFSnnn installed NFS(tm) Utilities (Network-Server/Communications) If the NFS subset is not installed or is corrupt, install it by using the setld command. See System Administration for more information on installing the subset.
	Check the /etc/rc.config file for the following entry: NFSSERVING="1" If the entry does not exist, run the NFS Configuration application. Be sure to specify that you will be exporting directories, even if you are going to edit the /etc/exports file manually. See Section 8.3 for more information. Verify that the network software has been configured. See the solution at [Network configured?] in Section 13.3.
	Verify that the portmap daemon is running. Enter the following command: #  ps -e | grep portmap If the portmap daemon is not running, stop and restart NFS, using the following commands: #  /sbin/init.d/nfs stop #  /sbin/init.d/nfs start If the portmap daemon does not start, reboot the server.
	Verify that the NFS daemons are registered with the portmap daemon. Enter the following commands: #  rpcinfo -u server_name mount #  rpcinfo -u server_name nfs If neither is registered, start NFS by using the following command: #  /sbin/init.d/nfs start
	To verify that the NFS daemons are running, complete the following steps: Verify that a mountd process is running. Enter the following command: #  ps -e | grep mountd If a mountd process is running, go to step 2. If no mountd process is running, stop and start NFS by using the following commands: #  /sbin/init.d/nfs stop #  /sbin/init.d/nfs start Verify that a nfsd process is running. Enter the following command: #  ps -e | grep nfsd If no nfsd process is running, stop and start NFS by using the following commands: #  /sbin/init.d/nfs stop #  /sbin/init.d/nfs start
Problem still exists? Report your problem to Digital. See Chapter 16.	To verify that the files are being exported, complete the following steps: Verify that file is being exported. Enter the following command: #  showmount -e If the file is being exported, go to step 3. If the file is not being exported, check that the file has an entry in the /etc/exports file. If there is no entry in the /etc/exports file, edit the file and create an entry. Have the remote system mount the file. If the file is being exported and the users cannot mount the file, check the /etc/rc.config file to see if they are allowed to mount the file. Enter the following command: #  /usr/sbin/rcmgr get NONROOTMOUNTS If the NONROOTMOUNTS parameter is 0, only users running as root can mount files from this server. To allow users not running as root to mount the files, enter the following command: #  /usr/sbin/rcmgr set NONROOTMOUNTS 1 Verify that the mountd daemon is running with Internet address checking on. Enter the following command: #  ps -e | grep mountd If the -i option appears, the client's name and address must be in the /etc/hosts file, or in the BIND or NIS hosts database. Only known hosts can mount the file system. If the -d or -s option appears, the client system must be in the same BIND domain or subdomain, respectively, as the server. If the mountd daemon is returning stale file handles for exported files, send a hangup signal (SIGHUP) to the mountd daemon to force it to reread the /etc/exports file. Enter the following commands: #  ps -e | grep mountd #  kill -1 mountd_pid

13.9    Solving NFS Client Problems

	Verify whether the NFS subset is installed. Enter the following command: #  setld -i | grep OSFNFS The following message should appear: OSFNFSnnn installed NFS(tm) Utilities (Network-Server/Communications) If the NFS subset is not installed or is corrupt, install it by using the setld command. See System Administration for more information on installing the subset.
	Check the /etc/rc.config file for the following entry: NFS_CONFIGURED="1" If the entry does not exist, run the NFS Configuration application. See Section 8.3 for more information. Verify that the network software has been configured. See the solution for [Network configured?] in Section 13.3.
	Verify that the portmap daemon is running. Enter the following command: #  ps -e | grep portmap If you do not find the portmap daemon, stop and restart NFS, using the following commands: #  /sbin/init.d/nfs stop #  /sbin/init.d/nfs start If the portmap daemon does not start, reboot the client.
	If the client cannot mount a remote file system or directory, complete the following steps: If an error message appears on the user's terminal, see Appendix E for the error message and a description. Verify that the remote NFS server is on your local network and in your hosts database. Verify that the server daemons on the remote system are running. Enter the following command: #  rpcinfo -p server_name Verify that the server is exporting the files you are expecting. Enter the following command: #  showmount -e server_name See the solutions for solving NFS server problems in Section 13.8. If the server is running and you still have problems, check the Ethernet connections and the Internet connections between the client system and the remote server. Check with other clients on the network to determine if they are having problems with the remote server. Verify the mount command line or the entry in the /etc/fstab file, and check the following: The host name matches the name of the remote NFS server. The mount point exists on your system. If you get an authentication error, check the following: If you are not a superuser, the server allows nonroot mounts. Your host name is in the server's hosts database. If your system is not in the same domain as the server, the server performs domain checking. See mountd(8) for more information on server options.
Problem still exists? Report your problem to Digital. See Chapter 16.	If application programs that perform file-related tasks do not complete their tasks or take a long time in doing so, complete the following steps: If an error message appears on the user's terminal, see Appendix E for the error message and a description. Verify that the server is running. See the solutions for solving NFS server problems in Section 13.8. If the server is running, check that the nfsd daemon is accumulating CPU time. If it is not, kill it and restart it. If this does not solve the problem, reboot the server. If the remote file systems or directories are mounted with the hard option, the program continues when the server is running once again. Check with other clients on the network to determine if they are having problems with the remote server. If they are not, check the Ethernet connections and the internet connections between the client system and the remote server. Check whether any nfsiod daemons are running. Enter the following command: #  ps -e | grep nfsiod If no nfsiod daemons are running, start some. Enter the following command: #  /usr/sbin/nfsiod 7 Although the nfsiod daemons are not necessary for a client, they perform read-ahead and write-behind functions, which might make I/O faster. If file access requests succeed but file locking requests hang indefinitely, verify that the local rpc.statd and rpc.lockd daemons are running. Enter the following commands: #  ps -e | grep rpc.statd #  ps -e | grep rpc.lockd If they are not running, start them. Enter the following commands: #  /usr/sbin/rpc.statd #  /usr/sbin/rpc.lockd Also, verify that the local rpc.statd and rpc.lockd daemons are running on the server. Enter the following commands: #  rpcinfo -p server_name | grep status #  rpcinfo -p server_name | grep lockmgr If they are not running, contact the server system administrator.

13.10    Solving UUCP Problems

	Verify whether the uucp subset is installed. Enter the following command: #  setld -i | grep OSFUUCP The following message should appear: OSFUUCPnnn installed UNIX(tm)-to-UNIX(tm)Copy Facility (General Applications) If the uucp subset is not installed, install it by using the setld command. See System Administration for more information on installing the subset.
	Verify whether the Basic Networking Services subset (containing the tip and cu utilities) is installed. Enter the following command: #  setld -i | grep OSFCLINET The following message should appear: OSFCLINETnnn installed Basic Networking Services (Network-Server/Communications) If the Basic Networking Services subset is not installed, install it by using the setld command. See System Administration for more information on installing the subset.
	Check for entries in the Permissions, Devices, and Systems files in the /usr/lib/uucp directory. If there are no entries, run the uucpsetup script. See Section 9.3 for more information.
	Configure the network hardware as follows: Direct connections to remote host -- Use a null modem or modem eliminator cable to connect your system to the remote host. Phone line connection to remote host -- Use a cable to connect your system to a modem and another cable to connect your modem to a phone line. The modem you use must be compatible with the modem at the remote host. Make sure the modem is configured as follows: Forced data set ready (DSR) is disabled. Full or verbose status messages are enabled. Character echo is disabled. Use 8-bit characters with no parity. XON/XOFF flow control is disabled. TCP/IP connection to remote host -- Use a cable to connect your system to the network. Then, run the Network Configuration application to configure the network. See Section 2.3 for more information on setting up the network.
	If you cannot dial up the remote system, check the following: Make sure that the setup parameters (such as speed, parity, modem control, flow control, and other terminal characteristics) on the local and remote ends are properly defined for the type of modem you have. Dial the number to the remote node. If you do not get an "Attached" message or a login prompt, plug a telephone handset into the local telephone line to check for a dial tone. If you do not hear a dial tone, call you local carrier to fix this problem. If you get no message, check the cabling between the local system and the modem. If you get a dial tone, check that your modem is operational and perform diagnostic tests on your modem. See the modem manual for more information. From another handset, dial the local telephone line. If the local telephone rings and you can carry on a conversation, the telephone line on the local end is good. If you cannot pass voice traffic, or if there is no ring, call your local carrier to fix this problem. Repeat steps 2 and 3 on the remote node to resolve problems with the remote end. If the telephone line is operational, verify that the remote modem is set up to automatically answer incoming calls when the system raises the data terminal ready (DTR) signal. The system raises the DTR signal by issuing a uugetty or getty command on the port.
	Run the uucp tests to test the connection to the remote system. See Section 14.5 and Section 14.6.
Problem still exists? Report your problem to Digital. See Chapter 16.	If the tip command does not execute successfully, complete the following steps: Check that the system name, connection speed, and phone number are in the /etc/remote file or that the system name and connection speed are in the /etc/remote file and the phone number is in the /etc/phones file. See remote(4) and phones(4) for more information. Check the at entry in the /etc/remote file. If the entry is correct, create an entry for the modem in the /etc/acucap file. See acucap(4) for more information. Check that the remote system is configured to answer incoming calls.

13.11    Solving NTP Problems

	Check the /etc/rc.config file for the following entry: XNTPD_CONF="YES" If the entry does not exist, run the ntpsetup script. See Section 10.3 for more information.
	Verify that a xntpd process is running. Enter the following command: #  ps -e | grep xntpd If no xntpd process is running, start NTP by using the following command: #  /sbin/init.d/xntpd start
	If the ntpq or xntpdc command cannot find the server host, the following message appears: ***Can't find host hostname The hostname is not in the /etc/hosts file, the BIND hosts database, or the NIS hosts database. Edit the /etc/hosts file and add an entry for the server host.
	If you run one of the monitor programs and in the output from the peers command the reach column contains zeros (0s), complete the following steps: Contact the system administrator of the server and verify which NTP daemon the server is running. The entry for the server in the /etc/ntp.conf file must contain the phrase version x after the server name, as follows: server host1 version x Check the /etc/hosts file and verify that there is an entry for each NTP server specified in the /etc/ntp.conf file. If you are using either BIND or NIS for host information, verify that the hosts database has an entry for each NTP server. If the xntpdc hostname command does not display any information, check whether the hostname server is running NTP."
Problem still exists? Report your problem to Digital. See Chapter 16.	If the ntpq or xntpdc request times out, the following message appears: hostname: timed out, nothing received ***Request timed out Complete the following steps: The hostname is not running the xntpd daemon. Check with the system administrator for that system. The network connection has gone down. See the solutions for [Host reachable?] at the beginning of this chapter. If you still cannot solve the problem, complete the following steps: Check the /etc/rc.config file to make sure it contains an entry similar to the following: XNTPD_CONF="YES" export XNTPD_CONF XNTP_SERV1=server1 export XNTP_SERV1 XNTP_SERV2=server2 export XNTP_SERV2 XNTP_SERV3=server3 export XNTP_SERV3 XNTPD_OPTS="-g" export XNTPD_OPTS If this entry does not exist or is incorrect, run the /usr/sbin/ntpsetup script. See Section 10.3 for more information. Check the /etc/ntp.conf file and make sure the information in it is accurate. It should contain entries for hosts running NTP with whom you want to synchronize system time. Make sure the correct version number is specified for each server and peer. To correct any entries, run the /usr/sbin/ntpsetup shell script. See Section 10.3 for information on running the script. Check the daemon.log file in the /var/adm/syslog.dated/date directory for information about NTP problems on the system.

13.12    Solving SLIP Problems

	Verify that the correct number of Serial Line Internet Protocol (SLIP) pseudodevices are supported in the kernel by using the netstat -in command. If SLIP is supported, output similar to the following appears for each interface: sl0* 296 <Link> 0 0 0 0 0 The sl prefix indicates that SLIP is supported on the system. In this example there is one SLIP interface. If you need additional SLIP interfaces, specify them by adding the nslip=x atribute under the net: subsystem in the /etc/sysconfigtab file. See System Administration for information on adding more SLIP interfaces. On systems with 24 megabytes of memory, SLIP is not configured into the kernel. To add SLIP into the kernel, edit the system configuration file (/usr/sys/confhostname ) and add the following entry: options SL See System Administration for more information.
	Configure the network hardware as follows: Verify that you are using the correct hardware. See Section 4.1.2.1 for more information. Make sure the modem is configured as follows: Use 8-bit characters with no parity. Software flow control (XON/XOFF) is disabled. For dial-in systems, follow the guidelines in Section 4.1.3.1. For dial-out systems, follow the guidelines in Section 4.1.3.2.
	If a remote system cannot dial in to your system successfully, complete the following steps: Edit the /etc/slhosts file and include the debug option in the login entry that cannot log in. See slhosts(4) for more information. Instruct the remote user to dial in again. Check the daemon.log file in the /var/adm/syslog.dated/date directory for information on SLIP problems on the dial-in system.
	If you cannot dial out to the remote system, complete the following steps: Verify that the modem is working correctly: Edit the /etc/acucap file and include the db option in your modem's entry. This prints usefule information for debugging a new entry. See acucap(4) for more information. Verify SLIP setup. Do the following: Edit the startslip dial-out script file and specify the debug subcommand and a debug log file. Try to dial out again. Check the debug log file for information about SLIP dial-out problems.
	If you cannot communicate with the remote host and none of the debug messages shows an error, complete the following steps: Check that the IP addresses and netmasks are correct on both ends of the connection. Check the following SLIP configuration parameters at each end of the connection: Internet Control Message Protocol (ICMP) traffic suppression -- If enabled at either end of the connection, the ping command will fail. TCP header compression -- If enabled at one end, TCP header compression must be enabled or autoenabled on the other end.
	If you can communicate with the remote host but not the the network connected the the remote host, complete the following steps: If your local system is using the remote system as a gateway system, issue the netstat -rn command on the local system to verify that the remote SLIP address is the default gateway. On the gateway system (remote system), issue the iprsetup -d command to see if ipforwarding and ipgateway are on. On the gateway system, verify that gated is running. See gated(8) for more information.
Problem still exists? Report your problem to Digital. See Chapter 16.	If the startslip command does not complete successfully, complete the following steps: Build your kernel with the PACKETFILTER option. Use the tcpdump command to examine packets sent and received through the SLIP interface. See tcpdump(8) for more information.

13.13    Solving PPP Problems

	Verify that the Point-to-Point Protocol (PPP) is supported in the kernel by using the netstat -i command. If PPP is supported, output similar to the following appears: ppp0* 296 <Link> 0 0 0 0 0 ppp1* 296 <Link> 0 0 0 0 0 The ppp prefix indicates that PPP is supported on the system. In this example there are two PPP pseudodevices. If PPP is not supported, add PPP into the /etc/sysconfigtab file. See System Administration for information on adding PPP into the sysconfigtab file.
	Verify that the PPP interface is configured with the local and remote IP addresses and a network mask. Enter the following command: #  ifconfig pppx The x variable is the interface number (0, 1, 2, ...). If the PPP interface is configured, output similar to the following appears: ppp0: flags=10<POINTOPOINT>
	Configure the network hardware as follows: Direct connections to remote host -- Use a null modem or modem eliminator cable to connect your system to the remote host. Phone line connection to remote host -- Use a cable to connect your system to a modem and another cable to connect your modem to a phone line. The modem you use must be compatible with the modem at the remote host. Make sure the modem is configured as follows: Use 8-bit characters with no parity. All flow control is disabled.
	If you are logging messages to the console and the link comes up successfully, the following messages appear on the console: Local IP address: xx.xx.xx.xx Remote IP address: yy.yy.yy.yy If the link does not come up, check the following: Check that the serial connection is set up successfully. Use the chat -v command to log the characters the chat program sends and receives. Check that pppd starts on the remote system. Use the chat -v command to log the characters the chat program sends and receives. Check the PPP negotiation between the two peers. Use the pppd command with the debug option to log the contents of all control packets sent and received.
Problem still exists? Report your problem to Digital. See Chapter 16.	If network applications do not work successfully, this might indicate an IP-address assignment problem or a routing problem. Do the following: Use the ifconfig, netstat -i, netstat -r, ping, and traceroute commands. If you can communicate with the peer machine but cannot communicate with other other machines beyond that in the network, there is a routing problem. For instances where the local machine is connected to the Internet through the peer, do the following: Assign the local machine an IP address on the same subnet as the remote machine. Run the local pppd daemon with the defaultroute option. Run the remote pppd daemon with the proxyarp option.

13.14    Solving LAT Problems

	Verify whether the Local Area Transport subset is installed. Enter the following command: #  setld -i | grep OSFLAT The following message should appear: OSFLATnnn installed Local Area Transport (LAT) (General Applications) If the subset is not installed, install it by using the setld command. See System Administration for information on installing the subset.
	Verify whether the Local Area Transport is configured in the kernel. Enter the following command: #  sysconfig -q lat If nothing is displayed, LAT is not configured in the kernel. Reconfigure the kernel with the LAT option. See System Administration for information on reconfiguring the kernel.
	Check the /etc/rc.config file for the following entry: LAT_SETUP="1" If the entry does not exist, run the latsetup utility. See Section 5.3 for information.
	If latsetup fails while creating new LAT ttys, check that /usr/sbin is included in the search path. Enter the following command: #  echo $PATH If it is not, include it in your PATH environment variable. Then, create new LAT ttys using latsetup.
	Verify whether LAT has been started. Enter the following command: #  latcp -d If LAT has been started, the following line appears: LAT Protocol is active If LAT has not been started, start it. Enter the following command: #  latcp -s
	If LAT starts up and messages appear continually on the system console, check the following messages and perform the required steps: Message 1 getty: cannot open "/dev/lat/xx". errno: 2 This means a LAT terminal device file (tty) does not exist and the /etc/inittab file contains an entry for this file. The latsetup utility will also report that no LAT entries are available. Do the following: Edit the /etc/inittab file and remove the LAT getty entries. If LAT terminal devices are required, create the LAT terminal device files and corresponding entries in the /etc/inittab file by using the latsetup command. See latsetup(8) for information. Message 2 getty: cannot open "/dev/lat/xx". errno: 19 This means the kernel was not configured with the LAT option and the /etc/inittab file contains at least one LAT getty entry. Do either of the following: Configure LAT into the kernel. See System Administration for information on configuring LAT into the kernel. Remove the LAT getty entries from the /etc/inittab file, either manually or by using the latsetup command. Message 3 INIT: Command is respawning too rapidly. The following meanings are possible: You are using an optional service name, such as lattelnet, and it is incorrectly defined. Do the following: Verify that the optional service name defined by the latcp -A command is correct by using the latcp -d command. Edit the /etc/inittab file and verify that a LAT entry has the optional service name specified correctly. An attempt to use a nonexistent LAT terminal device (tty) was made. Do the following: Edit the /etc/inittab file and remove the entry that has the nonexistent terminal device name. If LAT terminal devices are required, create the LAT terminal device files and corresponding entries in the /etc/inittab file by using the latsetup command. See latsetup(8) for more information.
	If the user cannot connect to or display a service from a terminal server via LAT, complete the following steps on the Digital UNIX system: Check if the service name is correct, using the latcp -d command. If the service name is incorrect, delete the service with the incorrect name. Enter the following command: #  latcp -D -aservice_name Then, add a service with the correct name. Enter the following command: #  latcp -A -aservice_name See latcp(8) for more information. Display the group codes for the service to which the user is attempting to connect using the latcp -d command. Check whether any group code matches a group displayed by the show port command at the terminal server. If no group code matches, do either of the following: Add at least one group displayed by the port to the service. Enter the following command: #  latcp -glist -aservice_name Change the port characteristics at the terminal server by adding a group that matches the service. See latcp(8) for more information. Check whether LAT is started on the system. If it is not, start it. Enter the following command: #  latcp -s If the problem persists, restart LAT. Enter the following command: #  latcp -s
	If problems occur when using an optional service, complete the following steps: Check whether the service was added as an optional service. Enter the following command: #  latcp -d Look for the following line: Service name: name (Optional) If Optional is not displayed, the optional service was not defined with the -o option. Delete the service. Enter the following command: #  latcp -D -aservice_name Then, add the service with the correct name and -o option. Enter the following command: #  latcp -A -aservice_name -o See latcp(8) for more information. Check if the optional service name matches the name defined in the /etc/inittab file. If it does not, do either of the following: Edit the /etc/inittab file and specify the optional service name. Delete the service. Enter the following command: #  latcp -D -aservice_name Then, add the service with the correct name and -o option. Enter the following command: #  latcp -A -aservice_name -o See latcp(8) for more information.
	If the user cannot connect to a host using LAT, the following messages appear: Connection to node-name not established. Service in use. The /etc/inittab file does not contain a sufficient number of getty entries. Create more LAT terminal devices (ttys) and add their corresponding entries into the /etc/inittab file by using the latsetup command. Then, restart LAT to advertise the available services. Enter the following command: #  latcp -s See Section 5.3 for information.
	If a host-initiated connection fails, check that the port, host, and service names are specified correctly. Enter the following command: #  latcp -d -P -L If these names are not specified correctly, delete the application ports with the incorrect name. Enter the following command: #  latcp -D -pport_name Then, add the application ports, using correct spelling. Enter either of the following commands: #  latcp -A -plocal_port -Hnode -Rrem_port #  latcp -A -plocal_port -Hnode -Vsvc_name See latcp(8) for information. Note When you delete an application port for a LAT printer, any currently executing print operation continues until the printer buffer is empty. The print job might not be complete.
	If you print a file to a printer attached to a LAT application port, the printer is online, and no printing occurs, check the status of the print queue. Enter the following command: #  lpc status The following line might appear: waiting for printer to become ready (offline ? If this line appears, verify whether LAT has been started. Enter the following command: #  latcp -d If LAT has not been started, start it. Enter the following command: #  latcp -s
	If problems are encountered with LAT/Telnet gateway, check the syslogd daemon messages file. Use the error message to eliminate the error. See Section 14.8 for more information on viewing the daemon.log file. The lattelnet utility uses the syslog message priority of LOG_INFO. For example, if you edit a LAT terminal entry in the /etc/inittab file, reassign it to lattelnet while a getty process is still active for the terminal, and a user tries to connect to lattelnet, the connection will fail. The following error message appears in the daemon.log file: No such file or directory Terminate the getty process for the terminal port.
Problem still exists? Report your problem to Digital. See Chapter 16.	If the LAT connection terminates abnormally, complete the following steps: Check the LAT terminal device (ttys) files for duplicate minor numbers. Enter the following command: #  ls -l /dev/lat/* If any exist, remove the duplicate device files, leaving the original file. Check the /etc/inittab file for duplicate LAT entries. Remove the duplicate entries, leaving the original entry.

13.15    Solving sendmail Problems

	Check the /etc/rc.config file for the following entry: MAIL_CONF="YES" If the entry does not exist, run the Mail Configuration application. See Section 11.3 for more information.
	Verify whether sendmail has been started. Enter the following command: #  ps -e | grep sendmail If sendmail is not running, start it, using the following command: #  sendmail
	If a user cannot send mail to another user, complete the following steps: Check whether the aliases database was changed. If it was, update the database by using the newaliases command. Check the mail.log files generated by the syslogd daemon for the specific mail message. If the message reached its destination, the addressee is not on the destination system. Verify that the user has the correct address. See Section 14.8 for information on viewing the syslogd message files.
	If you sent a mail message and the recipient did not receive it, complete the following steps: Check whether the address is correct. Check whether the remote node is reachable by using the ping command. Look in the mail.log files generated by the syslogd daemon for the sender's user name. See Section 14.8 for information on viewing the syslogd message files. If you find an entry, write down the message ID. If no entry is found, send the message again. Using the message ID, search through the mail.log files for the "from" and "to" entries. If you find a "from" entry but no "to" entry, either sendmail did not receive the message or the message was corrupted. Check the /var/spool/mqueue directory for files containing the message ID by entering the following command: #  ls -l /var/spool/mqueue/*fAAmessage_ID If the control file (qfAAmessage_ID) is present but no data file (dfAAmessage_ID) is present, the message was lost. If you find both a "from" entry and a "to" entry and the status is deferred, the message is in the queue. If there is no corresponding sent entry, use the mailq command to send the file and report the reason for the deferral. If you find both a "from" entry and a "to" entry and the status is sent, the message was delivered. If a local delivery, the message reached the destination. If a remote delivery, have the system administrator on the remote host search for the message.
Problem still exists? Report your problem to Digital. See Chapter 16.	If sendmail is not working correctly, complete the following steps: Check for an error message in the rejected message. Check for error messages in the mail.log files generated by the syslogd daemon. See Section 14.8 for information on viewing the syslogd message file. See Appendix G for a list of sendmail error messages.