Merge "Fix documentation for usage of Nova Cells (v2)"

This commit is contained in:
Zuul 2019-07-23 06:56:53 +00:00 committed by Gerrit Code Review
commit 34173da672

View File

@ -197,8 +197,8 @@ segregation method currently provided by OpenStack Compute.
- Availability zones
- Host aggregates
* - **Use**
- A single :term:`API endpoint` for compute, or you require a second
level of scheduling.
- Shard and scale compute deployment while maintaining a single
:term:`API endpoint`.
- Discrete regions with separate API endpoints and no coordination
between regions.
- Logical separation within your nova deployment for physical isolation
@ -212,8 +212,8 @@ segregation method currently provided by OpenStack Compute.
- A single-site cloud with equipment fed by separate power supplies.
- Scheduling to hosts with trusted hardware support.
* - **Overhead**
- Considered experimental. A new service, nova-cells. Each cell has a full
nova installation except nova-api.
- Each Cell contains instances of services with overlapping
functionality.
- A different API endpoint for every region. Each region has a full nova
installation.
- Configuration changes to ``nova.conf``.
@ -227,22 +227,20 @@ segregation method currently provided by OpenStack Compute.
Cells and regions
-----------------
OpenStack Compute cells are designed to allow running the cloud in a
`OpenStack Compute Cells <https://docs.openstack.org/nova/latest/user/cells.html>`_
are designed to allow running the cloud in a
distributed fashion without having to use more complicated technologies,
or be invasive to existing nova installations. Hosts in a cloud are
partitioned into groups called *cells*. Cells are configured in a tree.
The top-level cell ("API cell") has a host that runs the ``nova-api``
service, but no ``nova-compute`` services. Each child cell runs all of
the other typical ``nova-*`` services found in a regular installation,
except for the ``nova-api`` service. Each cell has its own message queue
and database service and also runs ``nova-cells``, which manages the
communication between the API cell and child cells.
partitioned into groups called *Cells*. Cells are configured in a tree.
The top-level Cell ("API cell") has a host that runs the ``nova-api``
service, but no ``nova-compute`` services. The ``nova-compute`` runs
in a child Cell.
Each Cell has its own message queue and database service.
This allows for a single API server being used to control access to
multiple cloud installations. Introducing a second level of scheduling
(the cell selection), in addition to the regular ``nova-scheduler``
selection of hosts, provides greater flexibility to control where
virtual machines are run.
multiple compute installations with the usage of multiple Cells.
See `Nova Cells V2 Layout <https://docs.openstack.org/nova/latest/user/cellsv2-layout.html>`_
for further documentation.
Unlike having a single API endpoint, regions have a separate API
endpoint per installation, allowing for a more discrete separation.