| Hey Edwards, that is a concern many arise, today, there is not any both answer and implementations that would allow you to setup such. One approach is to work with stateless instances - using an orchestration tool such a Puppet of Chef, which would take care by itself of the spawning and restoration. It supposes to relegate all the data to the orchestration tool. In case a node fails, or even an instance - redeploying would take less time. I myself encountered though some customers topology which don't fit such approach (critical databases, etc...) for such case, you can either prep. a sleeping instance in one another node or setup a data replication between X instances. Network-wise you can automate the floating ip allocation using a detection solution. I think that feature (auto respawn) would come up soon (guys feel free to correct me), since this is a real concern. Regards, Razique
Nuage & Co - Razique Mahroua
Le 21 nov. 2012 à 08:51, <[email protected]> a écrit :
|
_______________________________________________ Mailing list: https://launchpad.net/~openstack Post to : [email protected] Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
