SGI Techpubs Library

IRIX 6.2  »  Books  »  Administrative  »  
 Network License System Administration Guide
(document number: 007-2115-002 / published: 1993-09-07)    table of contents  |  additional info  |  download
find in page

Chapter 9. Troubleshooting
Prev   Next

Chapter 9. Troubleshooting

This chapter contains troubleshooting procedures that you can use if you are having trouble obtaining a concurrent license. It also describes various problem situations that can occur in a NetLS environment and what to do about them. The sections in this chapter are:

Additional sources of information about problems are:

  • The list of known bugs in the NetLS Execution Only Environment Release Notes.

  • The file /var/adm/SYSLOG (/usr/adm/SYSLOG on IRIX Release 5.0 and earlier), which may contain messages that can help you understand an error condition.

Basic Troubleshooting Procedure for Concurrent Licenses

If you are having trouble obtaining a concurrent license for a licensed product, the procedure below can be used to identify the source of your problem.

  1. Verify that the file /var/netls/nodelock (/usr/netls/nodelock on IRIX Release 5.0 and earlier) does not contain a license for the product you want to use. If you find a node-locked license, comment it out using the procedure in “Deleting a Node-locked License” in Chapter 6.

  2. On your system, install netls_eoe.sw.netls_eoe if it isn't installed already to give you the commands that you need to troubleshoot your problem.

  3. Confirm that there is a Global Location Broker on the network and that it is running by using the procedure in “Checking Global Location Brokers” in Chapter 7.

  4. Verify that the clock is within 12 hours of the Global Location Broker clocks. Reset incorrect clocks.

  5. Confirm that there is at least one Network License Server running in the network and that at least one of them has a license for the product you want to use by following the procedure in “Checking Network License Servers” in Chapter 7.

  6. Try stopping and restarting netlsd(1M) on the Network License Server that has the license you want to use. See “Stopping Daemons” in Chapter 7 and “Restarting Daemons” in Chapter 7.

  7. Verify that all Global Location Broker clocks are within 10 minutes of each other and manually merge their databases using the procedure in “Routine Maintenance of Global Location Brokers” in Chapter 7.

  8. Stop and restart all Global Location Brokers and Network License Servers in the network using the procedures in “Stopping Daemons” in Chapter 7 and “Restarting Daemons” in Chapter 7.

glbd Won't Run

If glbd(1M) won't start, it may be for these reasons:

  • You weren't superuser when you tried to start the daemon.

  • glbd may already be running.

  • The llb or glb chkconfig(1M) flags are “off”.

  • You may not be starting glbd properly. The first time you start glbd, you must give it the –create option with either the –first or –from option. Subsequently, glbd should be started with no options.

  • llbd may not be running. It must be running before glbd can run.

  • An error condition occurred that was reported in /var/adm/SYSLOG (/usr/adm/SYSLOG on IRIX Release 5.0 and earlier).

  • The GLB database may be corrupted.

The next two sections describe ways to restore or replace a corrupted Global Location Broker. If there is a Global Location Broker whose database you can use to restore the corrupted database (the best way to restore a corrupted database), see “Restoring a Corrupted GLB Database Using Another Global Location Broker” in this chapter. If there is no Global Location Broker you can use, see “Recreating a Corrupted GLB Database” in this chapter.

Restoring a Corrupted GLB Database Using Another Global Location Broker

To restore a corrupted GLB database using another Global Location Broker, follow this procedure:

  1. If possible, start llbd and glbd on the system with the corrupted database.

  2. If you were able to start the daemons, use the drm_admin(1M) command with the reset subcommand to delete the database (see “Using drm_admin reset” in Chapter 7).

  3. If the database is corrupted so badly that llbd and glbd cannot run, as superuser manually delete two database files. On IRIX Release 5.0.1 or later, give this command:

    rm /var/ncs/glb.p /var/ncs/glb.e

    On IRIX Release 5.0 or earlier, give this command:

    rm /usr/etc/ncs/glb.p /usr/etc/ncs/glb.e

  4. If there are several Global Location Brokers, inform them that this Global Location Broker is out of service. Start the drm_admin command:

    /usr/etc/drm_admin

  5. Give these subcommands once for each good Global Location Broker to inform it that the corrupt Global Location Broker is out of service:

    drm_admin: set -o glb -h ip:goodGLB 
drm_admin: purgerep ip:corruptGLB

    drm_admin issues a warning and asks for confirmation.

  6. Use “Setting Up a Global Location Broker Manually” in Chapter 4 to create a new Global Location Broker (use the name of a good Global Location Broker for other_GLB_hostname in step 8).

  7. Use the drm_admin merge_all command to ensure that all Global Location Brokers are consistent (see “Routine Maintenance of Global Location Brokers” in Chapter 7).

  8. Restart netlsd on each Network License Server with these commands:

    /etc/init.d/netls stop 
/etc/init.d/netls start

Recreating a Corrupted GLB Database

If the corrupted GLB database is the only one, you have two options. The options are:

  • Delete the corrupted database:

    1. On IRIX Release 5.0.1 or later, give this command:

      rm /var/ncs/glb.p /var/ncs/glb.e

    2. On IRIX Release 5.0 or earlier, give this command:

      rm /usr/etc/ncs/glb.p /usr/etc/ncs/glb.e

    3. Use “Setting Up a Global Location Broker Manually” in Chapter 4 to re-create an empty, fresh database.

    4. Stop and restart all Network License Servers (see “Stopping Daemons” in Chapter 7 and “Restarting Daemons” in Chapter 7).

  • Restore the database from backup:

    1. Restore the two database files /var/ncs/glb.e and /var/ncs/glb.p (or /usr/etc/ncs/glb.e and /usr/etc/ncs/glb.p for IRIX Release 5.0 and earlier) from a backup.

    2. Because the backup version may have obsolete information, use the clean subcommand of the lb_admin(1M) command to remove obsolete entries from the GLB database. See “Cleaning GLB Databases” in Chapter 7 for the procedure.

    3. Restart netlsd on each Network License Server with these commands:

      /etc/init.d/netls stop 
/etc/init.d/netls start

netlsd Won't Run

When netlsd won't run, it is usually for one of these reasons:

  • The llb or netls chkconfig flags are set to “off”.

  • The Global Location Broker(s) and Network License Server(s) aren't set up properly. See “Checking Global Location Brokers” in Chapter 7 and “Checking Network License Servers” in Chapter 7 for information on verifying that they are operating correctly.

  • An error condition occurred that was reported in /var/adm/SYSLOG (/usr/adm/SYSLOG on IRIX Release 5.0 and earlier).

  • The Concurrent License database is corrupted. If the netlsd daemon won't start, it may be because the Concurrent License database is corrupted. If netlsd was killed, the database might be in an inconsistent state.

Follow these steps to attempt to salvage a corrupted Concurrent License database, and if that fails, to remove it:

  1. Become superuser:

    /bin/su - 
Password:

  2. Start netlsd with the –o option:

    /usr/etc/netlsd -o &

  3. Use ps(1) to see if it started:

    ps -e | grep netlsd 
  3842 ttyq1 0:00 netlsd

    If you get a line of output like the one shown above, netlsd started and the problem is solved.

  4. If netlsd didn't start, the next step is to remove the Concurrent License database. This is a drastic step that requires you to reinstall every license that was in the database.

    To delete the Concurrent License database, give these commands on a system running IRIX Release 5.0.1 or later:

    /etc/init.d/netls stop 
/etc/init.d/nck stop 
rm /var/netls/*db*

    On a system running IRIX Release 5.0 or earlier, give these commands:

    /etc/init.d/netls stop 
/etc/init.d/nck stop 
rm /usr/netls/*db*

  5. To restart the daemons, see “Restarting Daemons” in Chapter 7 A new, empty Concurrent License database is created.

  6. Reinstall the licenses that were in the database or restore them from a backup.

Problems While Installing a Concurrent License

If ls_admin(1M) doesn't show the Network License Server you want to install a license on, or it doesn't allow you to add vendor or product information, some of the most likely causes are:

  • The Network License Server wasn't properly set up. See “Checking Network License Servers” in Chapter 7.

  • The system where you are using ls_admin can't communicate with the system that is the Network License Server. Try running ls_admin on the Network License Server system.

  • The sysinfo number for this system may not be the same as the sysinfo number that the license was issued for. To check this, give the command:

    /etc/sysinfo -s

    The result should be the same number that is listed for sysinfo on the license information sheet. If the sysinfo is “Any”, then the license works with any system.

  • You may have typed a password incorrectly.

  • You may have entered the wrong ID or password. Make sure that the vendor ID, the vendor password, and the license password are all correct.

  • The license you are entering may be a node-locked license rather than a concurrent license. Check the license type listed on the license information sheet.

  • There is more than one license password listed on the license information sheet, and the vendor password was not correctly copied from one Network License Server to another. See Chapter 5, “Installing Concurrent Licenses.”

Global Location Brokers That Don't Know About Each Other

If a network has two or more systems that are Global Location Brokers and the system(s) started second and later were not set up using the glbd –from option, then the Global Location Brokers may have different lists of Network License Servers. When an application tries to find a Network License Server, it may or may not find one because the Global Location Brokers may contain different lists of Network License Servers. One Network License Server may not have licenses available, but another Network License Server not on the list may. Each time you try to get a license, the application may try to get it from either one. If you bring up ls_admin several times and once in a while get a different list of Network License Servers in the Servers area, then this is probably the problem.

There are three possible solutions to this problem:

  • Add a /var/ncs/glb_site.txt file (/usr/etc/ncs/glb_site.txt on IRIX Release 5.0 and earlier) onto all of the systems that will be requesting licenses. See “Using glb_site.txt Files” in Chapter 7 and the glb_site(4) manual page for more information.

  • Make sure that all of the Global Location Brokers know about each other. To do this, follow these steps:

    1. Identify all of the Global Location Brokers.

    2. Confirm that the chkconfig flags are set properly: Global Location Brokers should have glb and llb “on”, and Network License Servers should have netls and llb “on”.

    3. Stop the daemons on each Global Location Broker and each Network License Server by giving these commands:

      /bin/su - 
Password: 
/etc/init.d/netls stop 
/etc/init.d/nck stop

    4. Follow the procedure in “Restarting All Daemons on a Network” in Chapter 7 to restart all of the daemons.

  • Form a license group to isolate your NetLS environment.

    Using a license group isolates users from Global Location Brokers that are not part of their NetLS environment. See “Creating a New License Group” in Chapter 7 for instructions on forming a license group.

Getting a License Is Very Slow

If getting a license is very slow, the GLB databases may contain obsolete entries. To remove the obsolete entries, use the procedure in “Cleaning GLB Databases” in Chapter 7.

Propagation Queues Overflow

When a Global Location Broker is stopped or unavailable for a long time, the propagation queues at other Global Location Brokers accumulate updates for propagation to the unreachable system. Eventually, this accumulation can overflow the propagation queues.

If you are unable to revive or reconnect the unreachable system, invoke drm_admin and use the purgerep subcommand to forcibly delete the Global Location Broker. If the Global Location Broker later becomes reachable, use the drm_admin reset command to stop and remove the database, then re-create the Global Location Broker using the procedure in “Setting Up a Global Location Broker Manually” in Chapter 4.

Prev Table of Contents Next
Chapter 8. Error Messages  Appendix A. NetLS Files 

Network License System Administration Guide
(document number: 007-2115-002 / published: 1993-09-07)    table of contents  |  additional info  |  download

    Front Matter
    About This Guide
    Chapter 1. NetLS Concepts
    Chapter 2. The Licensing Process
    Chapter 3. Installing Node-locked Licenses
    Chapter 4. Setting Up Global Location Brokers and Network License Servers
    Chapter 5. Installing Concurrent Licenses
    Chapter 6. License Administration Tasks
    Chapter 7. Global Location Broker and Network License Server Administration Tasks
    Chapter 8. Error Messages
    Chapter 9. Troubleshooting
    Appendix A. NetLS Files
    Glossary
    Index


home/search | what's new | help