Automic Workload Automation

 View Only
  • 1.  Port configuration for multiple AE instances

    Posted Nov 27, 2019 04:40 AM
    Hello everyone,

    Let's discuss how to configure AE 11.2 and AE 12.2 on a same machine.

    Our goal: Test our environment before upgrading it from 11.2 to 12.2. We're testing version Automic.Workload.Automation_12.2.4_2019-10-30.

    To isolate two different versions, we did this:
    • Installed 12.2 under dedicated OS (AIX) user
    • Use different system name for 12.2
    • Increased all port numbers by 100 for 12.2
      Communication processes: 2317-2321
      Primary worker process: 2370
      Worker processes: 2371-2379
      REST API: 8098
    • Use dedicated DB schema for 12.2
    • CAPKI not installed yet (everything runs on a same machine)
    • Services in SMD file use same name in both versions
    After DB Load (UCYBDBLD) of db/general/12.2/UC_UPD.TXT, there is just client 0.

    Service Manager starts OK. Underlying services start OK (with CLI tool ucybsmcl).

    Right after services are started, couple of connections appears in UC4 Service-Manager dialog window. Is this fine?

    Connection to client 0 with default credentials is working. Now, we shut down services and start them again via ucybsmcl. After this start, no connections appear in Service Manager.

    AWI still shows login dialog. When trying to connect to client 0 for a second time, following errors are shown:
    ctscm2d.conti.de:2317 - TimeoutException
    ctscm2d.conti.de:2318 - TimeoutException
    ctscm2d.conti.de:2319 - A remote host refused an attempted connect operation.
    ctscm2d.conti.de:2320 - A remote host refused an attempted connect operation.
    ctscm2d.conti.de:2321 - Connection with JCP succeeded, but no CPs are available.

    Do you have any idea what is going wrong there? Does the configuration of 12.2 environment look OK?

    (Configuration details or log entries can be provided from us.)

    Thank you for any help.


  • 2.  RE: Port configuration for multiple AE instances
    Best Answer

    Posted Nov 28, 2019 03:55 AM
    Dear Libor,

    The configuration you mentioned looks good. And if everything worked after the first start, there should not be a problem with the configuration.

    My experience is, when I stopped my engine on linux and started it again immediately, the processes crash. The reason for that was, that the ports were not released yet by the operating system. Maybe you have the same problem. It might take several minutes for the ports to be free again. I have no idea, how long you must wait to be sure that the ports are free again. You can check with a command like netstat (not sure if it is the same on AIX) which ports are occupied by which processes. To be sure to have free ports,  if you have the chance to, you can restart the server.


    ------------------------------
    Regards, Nicole
    ------------------------------