ansible/inventory/group_vars/autocloud-backend-stg

58 lines
1.7 KiB
Text

---
# Define resources for this group of hosts here.
lvm_size: 20000
mem_size: 6144
num_cpus: 4
# for systems that do not match the above - specify the same parameter in
# the host_vars/$hostname file
tcp_ports: [
# These four ports are for outbound fedmsg
3000, 3001, 3002, 3003,
]
fas_client_groups: sysadmin-noc,sysadmin-fedimg,sysadmin-releng
# These people get told when something goes wrong.
fedmsg_error_recipients:
- sysadmin-fedimg-members@fedoraproject.org
# These are consumed by a task in roles/fedmsg/base/main.yml
fedmsg_certs:
- service: shell
owner: root
group: sysadmin
can_send:
- logger.log
- service: autocloud
owner: root
group: fedmsg
can_send:
- autocloud.image
- autocloud.image.running
- autocloud.image.success
- autocloud.image.failed
- autocloud.image.queued
# For the MOTD
csi_security_category: Moderate
csi_primary_contact: Cloudmeisters - sysadmin-fedimg-members@fedoraproject.org
csi_purpose: Run the autocloud testing backend
csi_relationship: |
This runs one of the two autocloud testing backends
This host relies on:
- A postgres db server running on db01. The db is called 'autocloud'.
The two autocloud-backend nodes populate that database with results.
- fedmsg messages produced by koji indicating that new images were built.
- queries to koji directly to download that content.
Running locally we have:
- fedmsg-hub which receives the initial notification from koji via fedmsg.
- redis, which stores a work queue populated by fedmsg
- the autocloud daemon, which pulls work from the redis queue and uses tunir
to test images in vagrant.
There's a cronjob that kills vagrant every 10 minutes if it got stuck on the
last test.