Changelog
=========

1.7.0 (2015-07-01)
------------------

- Automatic Kibana configuration (#16002).
- Parallelize localconfig-kvm-init (#14920).


1.6.8 (2015-06-03)
------------------

- Implement late autosigning when VMs are brought up with race conditions.


1.6.7 (2015-06-02)
------------------

- Delete non-existent Ceph pools if they contain no images (#14997).
- Remove now outdated Qemu/KVM management code as this has all moved
  to fc.qemu.

1.6.6 (2015-04-10)
------------------

- Remove explicit nagios notification for technical contacts. (#14900)
- Fix spurious bug in update-iptables invocation.


1.6.5 (2015-03-20)
------------------

- Ignore missing pools when deleting volumes. (#14947)


1.6.4 (2015-03-13)
------------------

- Improve node deletion: delete snapshots of VM images, too. (#14894)


1.6.3 (2015-03-05)
------------------

- Improve node deletion: delete bacula config snippets, too.


1.6.2 (2015-03-02)
------------------

- Add "postpone" maintenance script status (#13755).
- Remove obsolete "redundancy_groups" action.


1.6.1 (2015-02-09)
------------------

- Fix bug for puppet master deactivation.


1.6 (2015-02-05)
----------------

- Use new deletions API to clean up VM data on puppetmaster, KVM hosts,
  Ceph, Bacula, and Nagios.

- Fix bogus ProcessLookupError except clause.


1.5.4 (2014-12-12)
------------------

- Fix bug which caused disk grow to fail during the first attempt (#13930).
- Improve iptables reloading (#13879).


1.5.3 (2014-11-22)
------------------

- Refactored resize2fs disk grow code (#13414).
- Delete expired snapshots which adhere to the '*-keep-until-*' naming
  convention (#13414).
- Remove old KVM configure code. The code lives now in the fc.qemu package
  (#13414).


1.5.2 (2014-10-01)
------------------

- Revive localconfig-resize2fs-root: this is not obsolete but necessary for
  disk grow.

- Exclude users from duplicate contact detection in case they have no email
  address.


1.5.1 (2014-10-01)
------------------

- Ensure we do not block restarts of VMs when the directory is not up.

- Ensure we screen directory exceptions when generating Nagios contacts.

- Remove old, unused, and broken KVM entry points.

- Avoid duplicates when creating technical contacts. (#13629)

- Explain how to run tests in HACKING.txt.


1.5.0 (2014-09-13)
------------------

- Integate new fc.qemu package instead of instrumenting a complicated
  init script.

- Restart firewall in case of changed config files.


1.4.1 (2014-09-09)
------------------

- Let Nagios configure contacts for technical contacts of resource groups, too.


1.4.0 (2014-08-22)
------------------

- Clean outdated RBD snapshots (#13414).


1.3.6 (2014-08-21)
------------------

- Bugfix: the actual config file we pass to qemu is the shell-expanded one in
  /run/kvm.*.cfg. The last refactoring accidentally pointed it to the template
  in /etc/kvm/.


1.3.5 (2014-08-21)
------------------

* Do not try to inmigrate VMs that should be running and *are* running already.

* Adapt test setup to py.test after moving to separate repository and
  get all tests green.

* Extract computation of tmp and swap volumes from init script so we can test
  them and make them more robust for migration.
