List of all important VMs in Windows CI

1. VMware:

   • ci-vc - vCenter Server on Windows
     • CRITICAL
   • ci-vc-um - vCenter Update Manager

2. Base OS templates:

   • Windows - Install OS (manually) + instruction (manually)
     • Preferrably: automate
   • CentOS - Prepared manually, no instruction
     • TODO: Instruction with requirements
     • Requirements:
       • root access SSH (password-based)
       • 2 NICs
       • TODO: Verify if there are no more requirements
   • Ubuntu - No template, required to provision Ubuntu-based VMs
     • TODO: Instruction/automation
     • Requirements:
       • ciadmin user (SSH accessible, password-based)
       • python3
       • 1 NIC
       • TODO: Verify if there are no more requirements

3. CI templates:

   • builder - Ansible, Jenkins job
     • Requires Windows template
   • tester - Ansible, Jenkins job
     • Requires Windows template
   • testbed - Ansible
     • Requires Windows template

4. Infra:

   • builders - Ansible, Jenkins job
   • testers - Ansible, Jenkins job
   • mgmt-dhcp - Created manually, DHCP server + configuration (172.17.0.0/24)
     • CRITICAL
     • impact if down:
       • will cause failure in all production CI jobs due to lack of dhcp for testbed machines
       • will not affect demo-env, dev-env and other machines in public network
     • fix cost:
       • Automate deployment
       • Before automation, backup
       • VMware HA
   • winci-drive - Created manually, network drive with (probably) temporary content
     • impact if down:
       • no impact on production CI
       • degrades debuggability of CI (cannot upload or use ready artifacts)
       • impacts ability to create builder templates
     • fix cost:
       • will need to recreate directory structure and copy contents from one of the builders
       • this process is not documented to this degree of detail (no list of dependencies)
       • Before automation, backup
       • VMware HA
   • winci-jenkins - Created manually, plugins + configuration
     • CRITICAL
     • impact if down:
       • ABSOLUTELY CRITICAL
     • fix cost:
       • BACKUP
         • job configuration
         • job logs
         • logs server ssh keys
         • jenkins plugin list (initial version on contrail-windows-ci; need to check if works)
         • credentials
       • Automate configuration
       • VMware HA
   • winci-mgmt - Created manually, Python requirements.txt have to be installed
     • CRITICAL
     • impact if down:
       • ABSOLUTELY CRITICAL (cannot create testenvs)
     • fix cost:
       • Automate provisioning
         • Ubuntu configured for Ansible
         • Ubuntu configured for Python tests
         • Ubuntu configured for monitoring scripts
       • Backup
         • ansible-vault-key
         • logs server ssh keys
         • github deploy ssh key
       • VMware HA
   • winci-registry – Created manually @ 10.84.12.27 – Docker registry with kolla and contrail-openstack-* images from ci-repo.englab.juniper.net:5000. We're mirroring, because we may use some older snapshot of images, that are already removed from ci-repo registry.
     • impact if down:
       • ABSOLUTELY CRITICAL (cannot create testenvs)
       • In some cases we can switch back to ci-repo.
     • fix cost:
       • Manual provisioning using provisioning manual
       • Backup
         • Images removed from ci-repo may be unavailable, another registry?
       • VMware HA
   • winci-vyos-mgmt
     • CRITICAL
     • impact if down:
       • ABSOLUTELY CRITICAL (cannot deploy contrail controllers)
     • fix cost:
       • Backup
       • Document configuration
       • VMware HA
   • winci-zuulv2-production - Created manually with Ansible script
     • CRITICAL
     • impact if down:
       • ABSOLUTELY CRITICAL (cannot run jobs)
     • fix cost:
       • Already automated
         • Create prod and dev inventories for Zuul
       • Backup:
         • gerrit ssh keys
         • jenkins ssh keys
       • VMware HA
   • winci-purgatory - Probably created manually? What is installed there?
     • CRITICAL
     • impact if down:
       • ABSOLUTELY CRITICAL (cannot run jobs)
     • fix cost:
       • pending PR for automation

5. Hosts:

   • esxi1-5
     • fix cost:
       • VMware HA on critical virtual machines
       • backup konfiguracji
   • SSD/SATA drivers
     • fix cost:
       • RAID on physical servers:
         • Not possible due to lack of hardware support
       • If not RAID, then shared storage for critical VMs is needed
         • Preferrably some storage array with RAID configured
   • NICs
     • we lack redundancy in case of main NIC failure

6. Backup:

   • Refer to Infrastructure Backups