v1.6.4

Hot fix release for PVE 6.x.

Changes/Fixes
91fc05f Patches the pvecm CLI library to fix cluster join failures (i.e. retain backwards compatibility). This change will be reverted once upstream has pushed an update with this patch. See #91 for more information.
#90 pvelocalhost is no longer inserted into system host files.
#89 pve_manage_ssh can now be used to toggle SSH configuration management. You can use this if you use a role to manage the entire SSH config file - but please note that cluster configuration may become a bit more difficult (as this role inserts config blocks based on your inventory).

v1.6.3

Bug fix release for PVE 6.x, and HA group configuration support

Changes/Fixes
#50 - HA groups can now be managed
#80 - PVE auth users are now properly created with specified passwords (and other attributes)
#81 - Idempotency checks for Ceph OSD creation is now more correct
#87 - Size attributes can now be defined for Ceph pools
#88 - Subscription patch no longer fails on PVE 6.x
a3720a9 - Vagrant deployment also brings up a Ceph storage cluster

v1.6.2

Minor bug fix release and Vagrantfile addition.

#76 Vagrant can now be used to bring up a local PVE cluster. It's also now possible to use Python 3 as the interpreter when deploying this role.
#75 ring0/link0 and related role variables have been replaced with pve_cluster_addr0/1
#74 pve_ceph_cluster_network can now be specified to use another network for backend traffic
#70 apt dist-upgrade is now used in lieu of apt upgrade when pve_run_system_upgrades is true
#69 Ceph improvements, namely one to help idempotency. (Ceph testers needed!)

DEPRECATION NOTICE

Support for the following role variables will be removed in the next minor version (1.7.0):

pve_cluster_ring0_addr
pve_cluster_ring1_addr
pve_cluster_bindnet0_addr
pve_cluster_bindnet1_addr
pve_cluster_link0_addr
pve_cluster_link1_addr

If you are overriding any of these, please update your playbooks to use pve_cluster_addr0 and/or pve_cluster_addr1 before updating to v1.7.0.

v1.6.1

Bug fix release for Ceph deployments on PVE 6/Debian Buster.

#63 - Introduces pve_ceph_repository_line for customizing the Ceph package repository used (and defaults to the correct one per version)
#64 - Introduces a pve_ssh_port that allows you to specify which SSH port that hosts in a cluster are listening on. This does not reconfigure SSH to listen on this port; rather, it configures the SSH client what port to use when connecting to other hosts in the cluster.

v1.6.0

This is a major feature release, bringing in support for PVE 6.x deployments.

Major changes:
#60 - PVE 6.x support
#52 - Ceph can now be deployed on PVE clusters
#51 - datacenter.cfg customization

Minor changes (bug fixes/enhancements): #57, #58, #48, #55

v1.5.0

This is a major feature release but includes a critical bug fix (that can't be backported).

Major changes:
#33 - PVE 4.x support has been removed
#38 - Fixes pvesh parsing
#45, #46 - Proxmox Storages configuration

v1.4.12

This is an important bug fix release. Please upgrade to this release to prevent issues adding nodes to a cluster or creating a new cluster.

42495cc Fixes add node tasks during cluster creation for PVE 5.x
0afc625 Preliminary support for Ansible 2.6 (current deprecation warnings resolved)
b010844 Testing environment refresh, with addition of tests for PVE 4.x on older Ansible versions

v1.4.11

This feature release includes the following:

#28 - Let's Encrypt SSL certificate support (see README for details)
#27 - apt-get update and apt-get dist-upgrade are no longer run by default on every run.
#26 - Limit root login between PVE cluster hosts to key authentication

The pve_upgrade variable has been deprecated in this release for the solution in #27. If you wish to use this role to keep installing system updates, please specify pve_run_system_upgrades: true in your playbook.

v1.4.10

Minor feature release.

By default, this role executes a apt-get dist-upgrade on every run. While this is required for initial installation, some users may not want this to occur when re-running their playbook after an initial installation run. You can configure this with the pve_upgrade role variable: no will disable all upgrades and safe will execute aptitude safe-upgrade (see apt_module docs for more information).

#25 - Allow upgrade step to be configurable

v1.4.9

Bug fix release for the subscription removal patch, as well as fixed a regression from 1.4.8 that removed enterprise repos for enterprise repo users.