Greetings everyone!
We in Ironic land would like to go into the PTG with some additional
thoughts, requirements, and ideas as it relates to distributed and
geographically distributed deployments.
As you may or may not know, we did take a first step towards
supporting some of the architectures needed with conductor_groups this
past cycle, but we have two very distinct needs that have been
expressed with-in the community.
1) Need to federate and share baremetal resources between ironic
deployments. A specification[1] was proposed to try and begin to
capture what this would look like ironic wise. At a high level, this
would look like an ironic node that actually consumes and remotely
manages a node via another ironic deployment. Largely this would be a
stand-alone user/admin user deployment cases, where hardware inventory
insight is needed.
2) Need to securely manage remote sites with different security
postures, while not exposing control-plane components as an attack
surface. Some early discussion into this would involve changing
Conductor/IPA communication flow[2], or at least supporting a
different model, and some sort of light weight intermediate middle-man
service that helps facilitate the local site management.
With that in mind, we would like to schedule a call for sometime next
week where we can kind of talk through and discuss these thoughts and
needs in real time in advance of the PTG so we can be even better
prepared. We are attempting to identify a time with a doodle[3].
Please select a time and date, so we can schedule something for next
week.
Thanks,
-Julia
[1]: https://review.openstack.org/#/c/560152/
[2]: https://review.openstack.org/212206
[3]: https://doodle.com/poll/y355wt97heffvp3m
_______________________________________________
OpenStack-operators mailing list
OpenStack-operators@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
We in Ironic land would like to go into the PTG with some additional
thoughts, requirements, and ideas as it relates to distributed and
geographically distributed deployments.
As you may or may not know, we did take a first step towards
supporting some of the architectures needed with conductor_groups this
past cycle, but we have two very distinct needs that have been
expressed with-in the community.
1) Need to federate and share baremetal resources between ironic
deployments. A specification[1] was proposed to try and begin to
capture what this would look like ironic wise. At a high level, this
would look like an ironic node that actually consumes and remotely
manages a node via another ironic deployment. Largely this would be a
stand-alone user/admin user deployment cases, where hardware inventory
insight is needed.
2) Need to securely manage remote sites with different security
postures, while not exposing control-plane components as an attack
surface. Some early discussion into this would involve changing
Conductor/IPA communication flow[2], or at least supporting a
different model, and some sort of light weight intermediate middle-man
service that helps facilitate the local site management.
With that in mind, we would like to schedule a call for sometime next
week where we can kind of talk through and discuss these thoughts and
needs in real time in advance of the PTG so we can be even better
prepared. We are attempting to identify a time with a doodle[3].
Please select a time and date, so we can schedule something for next
week.
Thanks,
-Julia
[1]: https://review.openstack.org/#/c/560152/
[2]: https://review.openstack.org/212206
[3]: https://doodle.com/poll/y355wt97heffvp3m
_______________________________________________
OpenStack-operators mailing list
OpenStack-operators@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators