Description of problem:
Scaling up large environments generates a lot of garbage that Director never purges/cleans up.
For example, customers report having have to clean up:
- old journal logs (journalctl --vacuum-time=1d) --> we gained around 2Gb
- old openstack projects logs
- old puppet report files ( /var/lib/puppet/reports/) --> we gained around 10Gb
(heat-manage purge_deleted does nothing)
Our recommendation is for dedicating at least 40 GB of disk space. But even with double that size, issues can occur in the long term due to running out of space (especially if it happens on upgrade/scaling operations).
Version-Release number of selected component (if applicable):
heat-manage purge_deleted is specifically there for for heat db cleanup and not journal or puppet report files.