Steps to reproduce the error:
1. Install devstack
2. enable domain-specific driver feature
domain_specific_drivers_enabled=true
domain_config_dir=/etc/keystone/domains
3. create an domain-specific conf file in /etc/keystone/domains/. (i.e. /etc/keystone/domains/keystone.acme.conf)
4. run 'keystone-manage domain_config_upload --domain-name acme' and you'll see a traceback similar to this
keystone-manage domain_config_upload --domain-name acme
4959 DEBUG keystone.notifications [-] Callback: `keystone.identity.core.Manager._domain_deleted` subscribed to event `identity.domain.deleted`. register_event_callback /opt/stack/keystone/keystone/notifications.py:292
4959 DEBUG oslo_db.sqlalchemy.session [-] MySQL server mode set to STRICT_TRANS_TABLES,STRICT_ALL_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,TRADITIONAL,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION _check_effective_sql_mode /usr/local/lib/python2.7/dist-packages/oslo_db/sqlalchemy/session.py:509
4959 CRITICAL keystone [-] AttributeError: 'CacheRegion' object has no attribute 'expiration_time'
4959 TRACE keystone Traceback (most recent call last):
4959 TRACE keystone File "/usr/local/bin/keystone-manage", line 6, in <module>
4959 TRACE keystone exec(compile(open(__file__).read(), __file__, 'exec'))
4959 TRACE keystone File "/opt/stack/keystone/bin/keystone-manage", line 44, in <module>
4959 TRACE keystone cli.main(argv=sys.argv, config_files=config_files)
4959 TRACE keystone File "/opt/stack/keystone/keystone/cli.py", line 600, in main
4959 TRACE keystone CONF.command.cmd_class.main()
4959 TRACE keystone File "/opt/stack/keystone/keystone/cli.py", line 543, in main
4959 TRACE keystone status = dcu.run()
4959 TRACE keystone File "/opt/stack/keystone/keystone/cli.py", line 513, in run
4959 TRACE keystone self.read_domain_configs_from_files()
4959 TRACE keystone File "/opt/stack/keystone/keystone/cli.py", line 481, in read_domain_configs_from_files
4959 TRACE keystone os.path.join(conf_dir, fname), domain_name)
4959 TRACE keystone File "/opt/stack/keystone/keystone/cli.py", line 399, in upload_config_to_database
4959 TRACE keystone self.resource_manager.get_domain_by_name(domain_name))
4959 TRACE keystone File "/usr/local/lib/python2.7/dist-packages/dogpile/cache/region.py", line 1040, in decorate
4959 TRACE keystone should_cache_fn)
4959 TRACE keystone File "/usr/local/lib/python2.7/dist-packages/dogpile/cache/region.py", line 629, in get_or_create
4959 TRACE keystone expiration_time = self.expiration_time
4959 TRACE keystone AttributeError: 'CacheRegion' object has no attribute 'expiration_time'
4959 TRACE keystone
So looks like keystone-manage CLI have a slimmed down version of load_backends
https:/
which doesn't have the cache configured
https:/
So we have two choices
1. add the cache configuration back to the slimmed down version. Drawback is that we have to perpetually maintaining two different versions of load_backends().
2. get rid of the slimmed down version and directly use the fat version in keystone/
But since keystone-manager CLI is something that will be used sparingly, I think we should go for option #2.