N1 Grid Service Provisioning System 5.0 Installation Guide

SSH Connectivity

Master Server Unable to Connect to Local Distributor Through an Intermediate Local Distributor

If the Master Server is unable to connect to another machine and displays a TTL expiry error after you use the Host Details page to update the configuration of that machine or any machine upstream, you might need to manually generate the transport.config file for some or all of the intermediate Local Distributors between that machine and the Master Server. Test the connection to each of the upstream Local Distributors of the problem machine by moving from the problem machine to the Master Server. For the Local Distributor to which you can successfully connect that is closest to the problem machine, regenerate the transport.config file and all of its downstream Local Distributors. Use the CLI Client net.gencfg command to generate transport.config files.

Unable to Connect to an Application Using SSH

If you are experiencing problems connecting to a machine after configuring the N1 Grid Service Provisioning System 5.0 to use SSH, follow the steps below to troubleshoot the problem.

ProcedureHow to Troubleshoot SSH Connectivity Issues

Before You Begin

If you are using ssh-agent, complete this task from the same session as the session that you used to start the ssh-agent.

Steps
  1. On the upstream machine, test the connection to the downstream machine.

    • To test the machine immediately downstream from the upstream machine, use the following command:


      # ssh target-IPaddress ls -l
      

      target-IPaddress is the IP address of the machine that is the furthest downstream that you want to test.

    • If you are using ssh-agent, to test a machine that is more than one other machine downstream from the machine on which you are running the ssh-agent, use the following command:


      # ssh -A target-IPadress-parentmachine
      ssh -A target-IPadress-parentmachine ssh -A target-IPaddress ls -l
      

      # ssh -A ssh -A target-machine-n-IPaddress ssh -A target-machine-2-IPaddress
      ssh -A target-machine-1-IPaddress ssh -A target-IPaddress ls -l
      

    target-machine-n-IPaddress are the IP addresses of the upstream Local Distributor machines of the machine being tested in the specified in order. For example, 1 is the machine that is closest to the machine being tested and n is the machine that is right before the Master Server. target-IPaddress is the IP address of the machine that is the furthest downstream that you want to test.

    target-IPadress-parentmachine is the IP address of any machine that is between the upstream machine and the downstream machine for which you are testing connectivity.

    If you are prompted for information, supply the information. Try the test again.

    If you are not prompted for information, continue to the next step.

  2. On the upstream machine, in the logger_config.xml file, before the <root> section, insert the following lines to enable logging with priority="debug":


    <category name="SSH.STDERR">
    <priority value="debug" />
    </category>
    <category name="com.raplix.rolloutexpress.net.transport.SshClientConnectionHandler">
    <priority value="debug" />
    </category>

    Wait for the upstream machine to read the log file updates.

  3. Test the connection again using the command that you used in Step 1.

    Examine the log output on the command line and in the SSH.STDERR log. Correct any problems found in the log files and try the test again.

    Examine the application log output on the upstream machine for the SSH command line you used to invoke the downstream application and the stderr output of the SSH command. Correct any problems identified by the logged messages and try the test again.

    If you do not find any problems in the log files, the upstream machine might be connecting properly to the downstream machine, but the application is not starting properly. Continue to the next step.

  4. Examine the ROX log file for errors starting the application on the downstream machine.

    • On Red Hat Linux and IBM AIX machines, examine the /tmp/ROXappnumbers.log file.

    • On Solaris OS machines, examine the /var/tmp/ROXappnumbers.log file.

    app is the application on the downstream machine that you are testing. Use Agent for a Remote Agent, Dist for a Local Distributor, and Proxy for a CLI Client. numbers are randomly generated numbers that are included in the file name.

  5. Correct any errors found in the log file.