infrastructure/ansible/host_vars/pve.yml

expose_ssh: true

private_ip: "{{ pve_hosts.pve.ip }}"
protected_ip: "{{ pve_hosts.pve.internal_ip }}"

zpools_to_scrub:
  - tank
  - rpool

# 10GB, or so
zfs_arc_size: 10000000000

sanoid_datasets:
  tank:
    use_template: production
    recursive: true
    process_children_only: true

  rpool:
    use_template: production
    recursive: true
Provision PVE and ingress VM 2020-12-21 17:11:38 +00:00			`expose_ssh: true`
Add ZFS configuration for PVE 2021-01-09 21:27:52 +00:00
Move telegraf to host This makes metric collection for SMART much simpler. I'll still be using the prometheus node exporter for actual system metrics, though. 2021-03-05 20:39:11 +00:00			`private_ip: "{{ pve_hosts.pve.ip }}"`
			`protected_ip: "{{ pve_hosts.pve.internal_ip }}"`

Add ZFS configuration for PVE 2021-01-09 21:27:52 +00:00			`zpools_to_scrub:`
			`- tank`
			`- rpool`
Limit ZFS ARC size on PVE 2021-01-09 22:32:55 +00:00
Increase ZFS RAM usage If i've got a load of RAM free, it might as well be being used to cache ZFS! 2021-03-06 21:38:21 +00:00			`# 10GB, or so`
			`zfs_arc_size: 10000000000`
Snapshot PVE root pool too 2021-01-10 13:23:36 +00:00
			`sanoid_datasets:`
			`tank:`
			`use_template: production`
			`recursive: true`
			`process_children_only: true`

			`rpool:`
			`use_template: production`
			`recursive: true`