
A new periodic task to automatically remove conductor records that have been offline for longer than a configured timeout period. This addresses the issue where deleted or decommissioned conductors would remain in the database indefinitely. Closes-Bug: #2069771 Assisted-by: Claude Sonnet 4.0 Change-Id: I90eb159abad94d8369b8792fa17c20d80201569a Signed-off-by: Afonne-CID <afonnepaulc@gmail.com>
33 lines
1.6 KiB
YAML
33 lines
1.6 KiB
YAML
---
|
|
features:
|
|
- |
|
|
Adds periodic cleanup of stale conductor entries from the database.
|
|
A new periodic task automatically removes conductor records that have
|
|
been offline for longer than the configured timeout period, helping
|
|
prevent accumulation of stale conductor entries.
|
|
|
|
Three new configuration options have been added to the ``[conductor]``
|
|
section:
|
|
|
|
* ``conductor_cleanup_interval`` - The interval in seconds of how often
|
|
to run the cleanup task (default: 86400 seconds - 1 day).
|
|
For example: 86400 seconds = 1 day, 604800 seconds = 1 week,
|
|
2592000 seconds = 30 days (approx. 1 month).
|
|
* ``conductor_cleanup_timeout`` - How long a conductor must be offline
|
|
before it's considered stale and eligible for cleanup
|
|
(default: 1209600 seconds - 2 weeks). This value is always required to be
|
|
at least 3x larger than ``[conductor]heartbeat_timeout`` since if
|
|
otherwise, active conductors might be mistakenly removed from the
|
|
database. The cleanup task will skip execution and log a warning if
|
|
this requirement is not met.
|
|
* ``conductor_cleanup_batch_size`` - Maximum number of stale conductors
|
|
to clean up in a single operation (default: 50).
|
|
|
|
fixes:
|
|
- |
|
|
[`Bug 2069771 <https://bugs.launchpad.net/ironic/+bug/2069771>`_]
|
|
Fixes an issue where deleted or decommissioned conductors would remain
|
|
in ``openstack baremetal conductor list`` indefinitely with
|
|
``Alive = False`` status. The new periodic cleanup task automatically
|
|
removes these stale entries after the configured timeout period.
|