[openstack-dev] [Heat] Using Job Queues for timeout ops

Joshua Harlow harlowja at outlook.com
Fri Nov 14 19:45:57 UTC 2014


Sounds like this is tooz[1] ;)

The api for tooz (soon to be an oslo library @ https://review.openstack.org/#/c/122439/) is around coordination and 'service-group' like behavior so I hope we don't have duplicates of this in 'oslo.healthcheck' instead of just using/contributing to tooz instead.

https://github.com/stackforge/tooz/blob/master/tooz/coordination.py#L63

CoordinationDriver
- watch_join_group
- unwatch_join_group
- join_group
- get_members
- ...

Tooz has backends that use [redis, zookeeper, memcache] to achieve the above API (it also has some locking support for distributed locks as well).

Feel free to jump on #openstack-state-management if u want more info (jd and the enovance guys and myself have developed that library for this kind of purpose).

-josh

On Nov 13, 2014, at 10:58 PM, Jastrzebski, Michal <michal.jastrzebski at intel.com> wrote:

> Also, on "Common approach to HA" session we moved something like oslo.healthcheck (or whatever it will be called), common lib for service-group like behavior. In my opinion it's pointless to implement zookeeper management in every project separately (its already in nova..). Might be worth looking closely into this topic.




More information about the OpenStack-dev mailing list