openstack
/
charm-nova-compute
Code Issues Proposed changes
charm-nova-compute/README.md

11 KiB
Raw Blame History

Overview

This charm provides Nova Compute, the OpenStack compute service. Its target platform is Ubuntu (preferably LTS) + OpenStack.

Usage

The following interfaces are provided:

  • cloud-compute - Used to relate (at least) with one or more of nova-cloud-controller, glance, ceph, cinder, mysql, ceilometer-agent, rabbitmq-server, neutron

  • nrpe-external-master - Used to generate Nagios checks.

Configuration

This section covers common and/or important configuration options. See file config.yaml for the full list of options, along with their descriptions and default values. See the Juju documentation for details on configuring applications.

openstack-origin

The openstack-origin option states the software sources. A common value is an OpenStack UCA release (e.g. 'cloud:xenial-queens' or 'cloud:bionic-ussuri'). See Ubuntu Cloud Archive. The underlying host's existing apt sources will be used if this option is not specified (this behaviour can be explicitly chosen by using the value of 'distro').

pool-type

The pool-type option dictates the Ceph storage pool type. See sections 'Ceph pool type' and 'RBD Nova images' for more information.

Ceph pool type

Ceph storage pools can be configured to ensure data resiliency either through replication or by erasure coding. This charm supports both types via the pool-type configuration option, which can take on the values of 'replicated' and 'erasure-coded'. The default value is 'replicated'.

For this charm, the pool type will be associated with Nova-managed images.

Note: Erasure-coded pools are supported starting with Ceph Luminous.

Replicated pools

Replicated pools use a simple replication strategy in which each written object is copied, in full, to multiple OSDs within the cluster.

The ceph-osd-replication-count option sets the replica count for any object stored within the 'nova' rbd pool. Increasing this value increases data resilience at the cost of consuming more real storage in the Ceph cluster. The default value is '3'.

Important: The ceph-osd-replication-count option must be set prior to adding the relation to the ceph-mon application. Otherwise, the pool's configuration will need to be set by interfacing with the cluster directly.

Erasure coded pools

Erasure coded pools use a technique that allows for the same resiliency as replicated pools, yet reduces the amount of space required. Written data is split into data chunks and error correction chunks, which are both distributed throughout the cluster.

Note: Erasure coded pools require more memory and CPU cycles than replicated pools do.

When using erasure coding two pools will be created: a replicated pool (for storing RBD metadata) and an erasure coded pool (for storing the data written into the RBD). The ceph-osd-replication-count configuration option only applies to the metadata (replicated) pool.

Erasure coded pools can be configured via options whose names begin with the ec- prefix.

Important: It is strongly recommended to tailor the ec-profile-k and ec-profile-m options to the needs of the given environment. These latter options have default values of '1' and '2' respectively, which result in the same space requirements as those of a replicated pool.

See Ceph Erasure Coding in the OpenStack Charms Deployment Guide for more information.

Ceph BlueStore compression

This charm supports BlueStore inline compression for its associated Ceph storage pool(s). The feature is enabled by assigning a compression mode via the bluestore-compression-mode configuration option. The default behaviour is to disable compression.

The efficiency of compression depends heavily on what type of data is stored in the pool and the charm provides a set of configuration options to fine tune the compression behaviour.

Note: BlueStore compression is supported starting with Ceph Mimic.

Database

Nova compute only requires database access if using nova-network. If using Neutron, no direct database access is required and the shared-db relation need not be added. The nova-network feature is not available in Ussuri and later, and so this interface produces a warning if added.

Networking

This charm support nova-network (legacy) and Neutron networking.

Ceph backed storage

This charm supports a number of different storage backends depending on your hypervisor type and storage relations.

RBD Nova images

To make Ceph the storage backend for Nova non-bootable disk images configuration option libvirt-image-backend must be set to 'rbd'. The below relation is also required:

juju add-relation nova-compute:ceph ceph-mon:client

RBD Cinder volumes

Starting with OpenStack Ocata, in order to maintain Cinder RBD support the below relation is required:

juju add-relation nova-compute:ceph-access cinder-ceph:ceph-access

This allows Nova to communicate with multiple Ceph backends using different cephx keys and user names.

Availability Zones

There are two options to provide default_availability_zone config for nova nodes:

  • default-availability-zone
  • customize-failure-domain

The order of precedence is as follows:

  1. Information from a Juju provider (JUJU_AVAILABILITY_ZONE) if customize-failure-domain is set to True and Juju has set the JUJU_AVAILABILITY_ZONE to a non-empty value;
  2. The value of default-availability-zone will be used if customize-failure-domain is set to True but no JUJU_AVAILABILITY_ZONE is provided via hook context by the Juju provider;
  3. Otherwise, the value of default-availability-zone charm option will be used.

The default_availability_zone in Nova affects scheduling if a given Nova node was not placed into an aggregate with an availability zone present as a property by an operator. Using customize-failure-domain is recommended as it provides AZ-aware scheduling out of the box if an operator specifies an AZ during instance creation.

These options also affect the AZ propagated down to networking subordinates which is useful for AZ-aware Neutron agent scheduling.

NFV support

This charm (in conjunction with the nova-cloud-controller and neutron-api charms) supports use of nova-compute nodes configured for use in Telco NFV deployments; specifically the following configuration options (yaml excerpt):

nova-compute:
  hugepages: 60%
  vcpu-pin-set: "^0,^2"
  reserved-host-memory: 1024
  pci-passthrough-whitelist: {"vendor_id":"1137","product_id":"0071","address":"*:0a:00.*","physical_network":"physnet1"}

In this example, compute nodes will be configured with 60% of available RAM for hugepage use (decreasing memory fragmentation in virtual machines, improving performance), and Nova will be configured to reserve CPU cores 0 and 2 and 1024M of RAM for host usage and use the supplied PCI device whitelist as PCI devices that as consumable by virtual machines, including any mapping to underlying provider network names (used for SR-IOV VF/PF port scheduling with Nova and Neutron's SR-IOV support).

The vcpu-pin-set configuration option is a comma-separated list of physical CPU numbers that virtual CPUs can be allocated to by default. Each element should be either a single CPU number, a range of CPU numbers, or a caret followed by a CPU number to be excluded from a previous range. For example:

vcpu-pin-set: "4-12,^8,15"

The pci-passthrough-whitelist configuration must be specified as follows:

A JSON dictionary which describe a whitelisted PCI device. It should take the following format:

["device_id": "<id>",] ["product_id": "<id>",]
["address": "[[[[<domain>]:]<bus>]:][<slot>][.[<function>]]" |
"devname": "PCI Device Name",]
{"tag": "<tag_value>",}

where '[' indicates zero or one occurrences, '{' indicates zero or multiple occurrences, and '|' mutually exclusive options. Note that any missing fields are automatically wildcarded. Valid examples are:

pci-passthrough-whitelist: {"devname":"eth0", "physical_network":"physnet"}

pci-passthrough-whitelist: {"address":"*:0a:00.*"}

pci-passthrough-whitelist: {"address":":0a:00.", "physical_network":"physnet1"}

pci-passthrough-whitelist: {"vendor_id":"1137", "product_id":"0071"}

pci-passthrough-whitelist: {"vendor_id":"1137", "product_id":"0071", "address": "0000:0a:00.1", "physical_network":"physnet1"}

The following is invalid, as it specifies mutually exclusive options:

pci-passthrough-whitelist: {"devname":"eth0", "physical_network":"physnet", "address":"*:0a:00.*"}

A JSON list of JSON dictionaries corresponding to the above format. For example:

pci-passthrough-whitelist: [{"product_id":"0001", "vendor_id":"8086"}, {"product_id":"0002", "vendor_id":"8086"}]

The OpenStack advanced networking documentation provides further details on whitelist configuration and how to create instances with Neutron ports wired to SR-IOV devices.

Network spaces

This charm supports the use of Juju network spaces (Juju v.2.0). This feature optionally allows specific types of the application's network traffic to be bound to subnets that the underlying hardware is connected to.

Note: Spaces must be configured in the backing cloud prior to deployment.

In addition this charm declares two extra-bindings:

  • internal: used to determine the network space to use for console access to instances.

  • migration: used to determine which network space should be used for live and cold migrations between hypervisors.

Note that the nova-cloud-controller application must have bindings to the same network spaces used for both 'internal' and 'migration' extra bindings.

Actions

This section lists Juju actions supported by the charm. Actions allow specific operations to be performed on a per-unit basis. To display action descriptions run juju actions ceph-mon. If the charm is not deployed then see file actions.yaml.

  • openstack-upgrade
  • pause
  • resume
  • hugepagereport
  • security-checklist

Bugs

Please report bugs on Launchpad.

For general charm questions refer to the OpenStack Charm Guide.