Follow the instructions below anytime prior to the event to setup the uber onsite server (typically, while it's running at the warehouse).  This step can be done anytime and it's recommended it be completed before going onsite, like a week or two before event.

These steps are all prerequisites for the next step, Migrating Ubersystem onsite from offsite, which happens typically the day before an event, after the entire network and internet has been installed onsite.

Devops Team

The steps below should be performed by (or working with) the devops team:

Setup a fresh server from scratch

  1. Get Techops to give Devops team root access to a freshly install ubuntu (todo: outdated version?) 14LTS VM and name it rams1.uber.magfest.org.
  2. Setup VPN with Tinc, and ssh access
    1. As root Use 'Tinc' to setup a VPN that allows our rams1 server to connect back to mcp.magfest.net
      1. TODO: document this . stuff is in /etc/tinc/ubernet . follow steps in this article for host named 'internalnyc': https://www.digitalocean.com/community/tutorials/how-to-install-tinc-and-set-up-a-basic-vpn-on-ubuntu-14-04
      2. TODO: make sure to share the new hosts/rams1 file into frontend.uber.magfest.org so it can SSH in and do backups
      3. the IP of rams1 on the VPN should be 192.168.0.3
    2. Setup SSH keys from mcp → rams1 for passwordless login under whichever user magbot runs as
    3. Setup SSH keys from frontend (our badly named backups server) → rams1 for passwordless login so we can do backups
    4. Test: Make sure you can login without a password by running:
      1. from mcp, as the user running magbot: ssh [email protected]
      2. from frontend, as the user running the backups: ssh [email protected]
  3. On MCP, ensure /etc/hosts has the following in it: 192.168.0.3 rams1.uber.vpn192.168.0.3 rams1.uber.magfest.org

Setup Magbot so it knows about the new onsite server

  1. on mcp, edit /home/dom/sysadmin/magbot/scripts/mag-deploy.coffee, add (or re-enable) the following:

prod_super_onsite: { host: "[rams1.uber.magfest.org](<http://rams1.uber.magfest.org/>)" # overridden in /etc/hosts action_type: "puppet_deploy" ssh_user: "root" role_needed: "uber_deploy"}, 2. in slack, issue command magbot die which will cause the bot to die and restart automatically with the new config changes 3. verify you can see prod_super_onsite with command magbot deploy list targets

Run deploy for the first time

This will setup an ubersystem deploy on rams1 WITHOUT MOVING ANY PRODUCTION DATA ON IT.  Later, once onsite we will move the production data from the cloud production server to rams1 as part of the migration.

Be aware that this server, while it doesn't contain any production data, is configured in non-dev mode and will attempt to send emails.  So be careful about adding any test attendees with real information here.

  1. magbot command**:magbot deploy to prod_super_onsite init_server production prime** (this command takes about 10 minutes to run)
  2. browse to https://rams1.uber.magfest.org/uber/ and you should see a working, but empty, setup (note: this will probably only work from inside the warehouse network, or if you're ssh'd into mcp.magfest.net)

Techops Team

Network setup

This step may be able to happen in pre-event setup, or we may have to wait until we're onsite.  Either way, make sure to test it before performing the onsite data migration.