openstack/kolla-ansible
Code Issues Proposed changes
5405cfecb9
kolla-ansible/doc/source/admin/production-architecture-guide.rst
Radosław Piliszek 277675ede0 Docs: Add IPv6 control plane (address families) 
IPv6 control plane implementation [1] follow-up.

[1] Ia34e6916ea4f99e9522cd2ddde03a0a4776f7e2c

Change-Id: Icc25463320c23fd510073bff0a8144437a3607a6
2019-10-23 10:10:38 +00:00

8.9 KiB
Raw Blame History

Production architecture guide

This guide will help with configuring Kolla to suit production needs. It is meant to answer some questions regarding basic configuration options that Kolla requires. This document also contains other useful pointers.

Node types and services running on them

A basic Kolla inventory consists of several types of nodes, known in Ansible as groups.

  • Control - Cloud controller nodes which host control services like APIs and databases. This group should have odd number of nodes for quorum.
  • Network - Network nodes host Neutron agents along with haproxy / keepalived. These nodes will have a floating ip defined in kolla_internal_vip_address.
  • Compute - Compute nodes for compute services. This is where guest VMs live.
  • Storage - Storage nodes, for cinder-volume, LVM or ceph-osd.
  • Monitoring - Monitor nodes which host monitoring services.

Network configuration

Interface configuration

In Kolla operators should configure following network interfaces:

  • network_interface - While it is not used on its own, this provides the required default for other interfaces below.
  • api_interface - This interface is used for the management network. The management network is the network OpenStack services uses to communicate to each other and the databases. There are known security risks here, so it's recommended to make this network internal, not accessible from outside. Defaults to network_interface.
  • kolla_external_vip_interface - This interface is public-facing one. It's used when you want HAProxy public endpoints to be exposed in different network than internal ones. It is mandatory to set this option when kolla_enable_tls_external is set to yes. Defaults to network_interface.
  • storage_interface - This is the interface that is used by virtual machines to communicate to Ceph. This can be heavily utilized so it's recommended to use a high speed network fabric. Defaults to network_interface.
  • cluster_interface - This is another interface used by Ceph. It's used for data replication. It can be heavily utilized also and if it becomes a bottleneck it can affect data consistency and performance of whole cluster. Defaults to network_interface.
  • swift_storage_interface - This interface is used by Swift for storage access traffic. This can be heavily utilized so it's recommended to use a high speed network fabric. Defaults to storage_interface.
  • swift_replication_interface - This interface is used by Swift for storage replication traffic. This can be heavily utilized so it's recommended to use a high speed network fabric. Defaults to swift_storage_interface.
  • tunnel_interface - This interface is used by Neutron for vm-to-vm traffic over tunneled networks (like VxLan). Defaults to network_interface.
  • neutron_external_interface - This interface is required by Neutron. Neutron will put br-ex on it. It will be used for flat networking as well as tagged vlan networks. Has to be set separately.
  • dns_interface - This interface is required by Designate and Bind9. Is used by public facing DNS requests and queries to bind9 and designate mDNS services. Defaults to network_interface.
  • bifrost_network_interface - This interface is required by Bifrost. Is used to provision bare metal cloud hosts, require L2 connectivity with the bare metal cloud hosts in order to provide DHCP leases with PXE boot options. Defaults to network_interface.

Warning

Ansible facts does not recognize interface names containing dashes, in example br-ex or bond-0 cannot be used because ansible will read them as br_ex and bond_0 respectively.

Address family configuration (IPv4/IPv6)

Starting with the Train release, Kolla Ansible allows operators to deploy the control plane using IPv6 instead of IPv4. Each Kolla Ansible network (as represented by interfaces) provides a choice of two address families. Both internal and external VIP addresses can be configured using an IPv6 address as well. IPv6 is tested on Debian and Ubuntu.

Warning

While Kolla Ansible Train requires Ansible 2.6 or later, IPv6 support requires Ansible 2.8 or later due to a bug: https://github.com/ansible/ansible/issues/63227

Note

Currently there is no dual stack support. IPv4 can be mixed with IPv6 only when on different networks. This constraint arises from services requiring common single address family addressing.

For example, network_address_family accepts either ipv4 or ipv6 as its value and defines the default address family for all networks just like network_interface defines the default interface. Analogically, api_adress_family changes the address family for the API network. Current listing of networks is available in globals.yml file.

Note

While IPv6 support introduced in Train is broad, some services are known not to work yet with IPv6 or have some known quirks:

Docker configuration

Because Docker is core dependency of Kolla, proper configuration of Docker can change the experience of Kolla significantly. Following section will highlight several Docker configuration details relevant to Kolla operators.

Storage driver

In certain distributions Docker storage driver defaults to devicemapper, which can heavily hit performance of builds and deploys. We suggest to use btrfs or aufs as driver. More details on which storage driver to use in Docker documentation.

Volumes

Kolla puts nearly all of persistent data in Docker volumes. These volumes are created in Docker working directory, which defaults to /var/lib/docker directory.

We recommend to ensure that this directory has enough space and is placed on fast disk as it will affect performance of builds, deploys as well as database commits and rabbitmq.

This becomes especially relevant when enable_central_logging and openstack_logging_debug are both set to true, as fully loaded 130 node cluster produced 30-50GB of logs daily.

High Availability (HA) and scalability

HA is an important topic in production systems. HA concerns itself with redundant instances of services so that the overall service can be provided with close-to-zero interruption in case of failure. Scalability often works hand-in-hand with HA to provide load sharing by the use of load balancers.

OpenStack services

Multinode Kolla Ansible deployments provide HA and scalability for services. OpenStack API endpoints are a prime example here: redundant haproxy instances provide HA with keepalived while the backends are also deployed redundantly to enable both HA and load balancing.

Other core services

The core non-OpenStack components required by most deployments: the SQL database provided by mariadb and message queue provided by rabbitmq are also deployed in a HA way. Care has to be taken, however, as unlike previously described services, these have more complex HA mechanisms. The reason for that is that they provide the central, persistent storage of information about the cloud that each other service assumes to have a consistent state (aka integrity). This assumption leads to the requirement of quorum establishment (look up the CAP theorem for greater insight).

Quorum needs a majority vote and hence deploying 2 instances of these do not provide (by default) any HA as a failure of one causes a failure of the other one. Hence the recommended number of instances is 3, where 1 node failure is acceptable. For scaling purposes and better resilience it is possible to use 5 nodes and have 2 failures acceptable. Note, however, that higher numbers usually provide no benefits due to amount of communication between quorum members themselves and the non-zero probability of the communication medium failure happening instead.