why does nova have so many tokens?

asked 2015-06-07 16:05:09 -0500

don gravatar image

updated 2015-06-07 16:08:26 -0500

juno, ubuntu 14.10.

If I run this query:

 mysql> select count(*) as count,user_id from token group by user_id order by count desc limit 5;
+-------+----------------------------------+
| count | user_id                          |
+-------+----------------------------------+
| 30400 | b9b6f116169145629090c92dd968fca4 |
| 10865 | 56801ce4202d46b88a0bd25d5d376965 |
|  1333 | 187aa5c19dbd44269c689f7d9ce28d92 |
|   991 | 27fa0c2b6238417bb7fbfbc02fcde980 |
+-------+----------------------------------+
5 rows in set (0.04 sec)

I see that b9b6f116169145629090c92dd968fca4 is nova, and 56801ce4202d46b88a0bd25d5d376965 is admin. This is over 10 days (token interval time).

Why does nova have so many tokens? I think(?) this relates to neutron creating a token per call back into nova, but that is just a guess. Does anyone know where to make what fix to make this use a cached token? Or is it from somewhere else?

I see some old threads on this, and a patch that is abandoned: https://review.openstack.org/#/c/49836/ but, is that it?

edit retag flag offensive close merge delete