nomad Directory Reference

Files
	wscript

Detailed Description

Nomad documentation for HLCC

Prerequisites

• CII Services configured and started
• OLDB datapoints initialized
• HLCC software and its dependencies installed form RPM or sources on the selected hosts
• Nomad and Consul cluster up and running:
  • The environment variable NOMAD_ADDR must be set accordingly with the proper IP address. For example, on the hlcc test machine, we have:

    NOMAD_ADDR=http://elthlccd63.hq.eso.org:4646

  • Nomad and Consul clusters have to be running. For example, on the hlcc test machine as root, we can execute:

        systemctl start consul
        systemctl start nomad
        @icode 
    * Selected client nodes with installed HLCC has [meta variables](https://developer.hashicorp.com/nomad/docs/configuration/client#meta):
      * `role` set to `hlcc`
      * `trs_health_enabled` set to `0` or `1`
      * `dds_profile` set to `Localhost_Only` or `Default`

    client { ... meta { ... role = "hlcc"

    # Set to true on machines with PTP configured and therefore # trs health shall be checked trs_health_enabled = "1"

    # set to # Localhost_Only on isolated machine, so that not DDS traffic comes in/out # Default on machines that should exchange DDS data with other machines dds_profile = "Default" } ... }

    * Selected client nodes with installed HLCC has `control_deterministic` and `control_non_deterministic`
      [host_networks](https://developer.hashicorp.com/nomad/docs/configuration/client#host_network-block) set. E.g.:

    client { host_network "control_deterministic" { cidr = "127.0.0.0/8" }

    host_network "control_non_deterministic" { cidr = "134.171.3.0/24" } }

    Usage
    -----
     
    The hlcc startup and shutdown sequences are responsible for starting and shutting down the nomad jobs.<br>
    It is anyway possible to start the jobs manually on the command line.<br>
    For example to start the main HLCC job:

    $ nomad job run <hlcc_installation_path>/nomad/hlcc.nomad.hcl

    i.e. on a standard RPM installation:

    $ nomad job run /elt/hlcc/resource/nomad/hlcc.nomad.hcl

    In the case of a source code installation in $INTROOT, it is necessary to let Nomad know where to find the files by passing the prefix variable.<br>
    The correct command line in this case is:

    nomad job run –var prefix=$INTROOT $INTROOT/resource/nomad/hlcc.nomad.hcl

    This and other corresponding files would be located:

    In an INTROOT installation in: $INTROOT/resource/nomad In the hlcc source tree in: hlcc/nomad/src

    Stopping HLCC on the command line:

    $ nomad job stop hlcc

    Commands to start/stop nomad jobs can be issued with the nomad command line tool from any machine on any cluster, assuming 

    NOMAD_ADDR is properly set.

HLCC and LSV jobs use following Input Variables:

• prefix (default: /elt/hlcc) - the path where HLCC software is installed
• config_template_infix (default: dev) - config template file name infix

Variables can be passed via the --var option i.e.:

nomad job run --var prefix=$INTROOT hlcc.nomad.hcl

Development

• Use two-space indent
• Double quotes are preferred for strings representation

Roadmap

• Test usage of dynamic blocks