Bug #20177
closed
Some notes (cribbed from email by Dan Gryniewicz):
The lifecycle config looks sane; Dan, do you see anything here?
objects are checked for expiration once per day. The start time for processing is configured by "rgw lifecycle work time", and defaults to midnight.
For the record, if you are just testing, we introduced a debug harness (used in Teuthology) that will make each "day" pass in a configurable number of seconds. This should not be used in any production setup, since the overhead can be quite high, but it can be used to avoid waiting a whole day when testing the feature. The configurable is "rgw lc debug interval"
is this still an ongoing issue?
- Subject changed from RGW lifecycle not expiring objects to RGW lifecycle not expiring objects due to permissions on lc pool
- Priority changed from Urgent to High
- Assignee changed from Yehuda Sadeh to Abhishek Lekshmanan
It is the same as the reshard error, we should init to namespace.
The same applys to the roles pool
- Status changed from New to In Progress
- Assignee changed from Abhishek Lekshmanan to Orit Wasserman
- Status changed from In Progress to Fix Under Review
- Status changed from Fix Under Review to Pending Backport
- Backport changed from c10.2.8 to luminous
- Copied to Backport #20967: luminous: RGW lifecycle not expiring objects due to permissions on lc pool added
- Status changed from Pending Backport to Resolved
Also available in: Atom
PDF