openstack/openstack-ansible-os_cinder
Code Issues Proposed changes
Files
4a9a00fb3b2f097cf1f5d2cf2d0883199157fc68
openstack-ansible-os_cinder/doc/source/configure-cinder.rst
Kourosh Vivan 4e7cb3441e Change ceph example backend name 
ceph pool name change from volumes_hdd to volumes to match ceph defaults
cinder backend name change from volumes_hdd to rbd_backend

Change-Id: Ia003f94b5266e3fa70f499d68139eff7b7fa7c6b
2017-10-26 17:17:36 +02:00

20 KiB
Raw Blame History

Configuring the Block (cinder) storage service (optional)

By default, the Block (cinder) storage service installs on the host itself using the LVM backend.

Note

While this is the default for cinder, using the LVM backend results in a Single Point of Failure.

The LVM back end needs to run on the host, however most of the other back ends can be deployed inside a container. If the storage back ends deployed within your environment are able to run inside containers, then it is recommended to set is_metal: False in the env.d/cinder.yml file.

Note

Due to a limitation of the container system, you must deploy the volume service directly onto the host when using back ends depending on iSCSI. That is the case, for example, for storage appliances configured to use the iSCSI protocol.

NFS backend

Edit /etc/openstack_deploy/openstack_user_config.yml and configure the NFS client on each storage node if the NetApp backend is configured to use an NFS storage protocol.

  1. For each storage node, add one cinder_backends block underneath the a new container_vars section. container_vars are used to allow container/host individualized configuration. Each cinder back end is defined with a unique key. For example, nfs-volume1. This later represents a unique cinder backend and volume type.

    container_vars:
  cinder_backends:
    nfs-volume1:

  2. Configure the appropriate cinder volume backend name:

    volume_backend_name: NFS_VOLUME1

  3. Configure the appropriate cinder NFS driver:

    volume_driver: cinder.volume.drivers.nfs.NfsDriver

  4. Configure the location of the file that lists shares available to the block storage service. This configuration file must include nfs_shares_config:

    nfs_shares_config: FILENAME_NFS_SHARES

    Replace FILENAME_NFS_SHARES with the location of the share configuration file. For example, /etc/cinder/nfs_shares_volume1.

  5. Define mount options for the NFS mount. For example:

    nfs_mount_options: "rsize=65535,wsize=65535,timeo=1200,actimeo=120"

  6. Configure one or more NFS shares:

    shares:
   - { ip: "HOSTNAME", share: "PATH_TO_NFS_VOLUME" }

    Replace HOSTNAME with the IP address or hostname of the NFS server, and the PATH_TO_NFS_VOLUME with the absolute path to an existing and accessible NFS share (excluding the IP address or hostname).

The following is a full configuration example of a cinder NFS backend named NFS1. The cinder playbooks will automatically add a custom volume-type and nfs-volume1 as in this example:

container_vars:
  cinder_backends:
    nfs-volume1:
      volume_backend_name: NFS_VOLUME1
      volume_driver: cinder.volume.drivers.nfs.NfsDriver
      nfs_shares_config: /etc/cinder/nfs_shares_volume1
      nfs_mount_options: "rsize=65535,wsize=65535,timeo=1200,actimeo=120"
      shares:
      - { ip: "1.2.3.4", share: "/vol1" }

Backup

You can configure cinder to backup volumes to Object Storage (swift). Enable the default configuration to back up volumes to a swift installation accessible within your environment. Alternatively, you can set cinder_service_backup_swift_url and other variables to back up to an external swift installation.

  1. Add or edit the following line in the /etc/openstack_deploy/user_variables.yml file and set the value to True:

    cinder_service_backup_program_enabled: True

  2. By default, cinder uses the access credentials of the user initiating the backup. Default values are set in the /opt/openstack-ansible/playbooks/roles/os_cinder/defaults/main.yml file. You can override those defaults by setting variables in /etc/openstack_deploy/user_variables.yml to change how cinder performs backups. Add and edit any of the following variables to the /etc/openstack_deploy/user_variables.yml file:

    ...
cinder_service_backup_swift_auth: per_user
# Options include 'per_user' or 'single_user'. We default to
# 'per_user' so that backups are saved to a user's swift
# account.
cinder_service_backup_swift_url:
# This is your swift storage url when using 'per_user', or keystone
# endpoint when using 'single_user'.  When using 'per_user', you
# can leave this as empty or as None to allow cinder-backup to
# obtain a storage url from environment.
cinder_service_backup_swift_url:
cinder_service_backup_swift_auth_version: 2
cinder_service_backup_swift_user:
cinder_service_backup_swift_tenant:
cinder_service_backup_swift_key:
cinder_service_backup_swift_container: volumebackups
cinder_service_backup_swift_object_size: 52428800
cinder_service_backup_swift_retry_attempts: 3
cinder_service_backup_swift_retry_backoff: 2
cinder_service_backup_compression_algorithm: zlib
cinder_service_backup_metadata_version: 2

During installation of cinder, the backup service is configured.

Using Ceph for cinder backups

You can deploy Ceph to hold cinder volume backups. To get started, set the cinder_service_backup_driver Ansible variable:

cinder_service_backup_driver: cinder.backup.drivers.ceph

Configure the Ceph user and the pool to use for backups. The defaults are shown here:

cinder_service_backup_ceph_user: cinder-backup
cinder_service_backup_ceph_pool: backups

Availability zones

Create multiple availability zones to manage cinder storage hosts. Edit the /etc/openstack_deploy/openstack_user_config.yml and /etc/openstack_deploy/user_variables.yml files to set up availability zones.

  1. For each cinder storage host, configure the availability zone under the container_vars stanza:

    cinder_storage_availability_zone: CINDERAZ

    Replace CINDERAZ with a suitable name. For example cinderAZ_2.

  2. If more than one availability zone is created, configure the default availability zone for all the hosts by creating a cinder_default_availability_zone in your /etc/openstack_deploy/user_variables.yml

    cinder_default_availability_zone: CINDERAZ_DEFAULT

    Replace CINDERAZ_DEFAULT with a suitable name. For example, cinderAZ_1. The default availability zone should be the same for all cinder hosts.

OpenStack Dashboard (horizon) configuration for cinder

You can configure variables to set the behavior for cinder volume management in OpenStack Dashboard (horizon). By default, no horizon configuration is set.

  1. The default destination availability zone is nova if you use multiple availability zones and cinder_default_availability_zone has no definition. Volume creation with horizon might fail if there is no availability zone named nova. Set cinder_default_availability_zone to an appropriate availability zone name so that Any availability zone works in horizon.
  2. horizon does not populate the volume type by default. On the new volume page, a request for the creation of a volume with the default parameters fails. Set cinder_default_volume_type so that a volume creation request without an explicit volume type succeeds.

Configuring cinder to use LVM

  1. List the container_vars that contain the storage options for the target host.

    Note

    The vars related to the cinder availability zone and the limit_container_types are optional.

    To configure an LVM, utilize the following example:

    storage_hosts:
 Infra01:
   ip: 172.29.236.16
   container_vars:
     cinder_storage_availability_zone: cinderAZ_1
     cinder_default_availability_zone: cinderAZ_1
     cinder_backends:
       lvm:
         volume_backend_name: LVM_iSCSI
         volume_driver: cinder.volume.drivers.lvm.LVMVolumeDriver
         volume_group: cinder-volumes
         iscsi_ip_address: "{{ cinder_storage_address }}"
       limit_container_types: cinder_volume

To use another backend in a container instead of bare metal, copy the env.d/cinder.yml to /etc/openstack_deploy/env.d/cinder.yml file and change the is_metal: true stanza under the cinder_volumes_container properties to is_metal: false.

Alternatively, you can also selectively override, like this:

container_skel:
  cinder_volumes_container:
    properties:
      is_metal: false

Configuring cinder to use Ceph

In order for cinder to use Ceph, it is necessary to configure for both the API and backend. When using any forms of network storage (iSCSI, NFS, Ceph) for cinder, the API containers can be considered as backend servers. A separate storage host is not required.

Copy the env.d/cinder.yml to /etc/openstack_deploy/env.d/cinder.yml file and change the is_metal: true stanza under the cinder_volumes_container properties to is_metal: false.

Alternatively, you can also selectively override, like this:

container_skel:
  cinder_volumes_container:
    properties:
      is_metal: false

  1. List of target hosts on which to deploy the cinder API. We recommend that a minimum of three target hosts are used for this service.

    storage-infra_hosts:
  infra1:
    ip: 172.29.236.101
  infra2:
    ip: 172.29.236.102
  infra3:
    ip: 172.29.236.103

    To configure an RBD backend, utilize the following example:

    container_vars:
  cinder_storage_availability_zone: cinderAZ_3
  cinder_default_availability_zone: cinderAZ_1
  cinder_backends:
    limit_container_types: cinder_volume
    rbd_backend:
      volume_driver: cinder.volume.drivers.rbd.RBDDriver
      rbd_pool: volumes
      rbd_ceph_conf: /etc/ceph/ceph.conf
      rbd_flatten_volume_from_snapshot: 'false'
      rbd_max_clone_depth: 5
      rbd_store_chunk_size: 4
      rados_connect_timeout: 30
      volume_backend_name: rbd_backend
      rbd_user: "{{ cinder_ceph_client }}"
      rbd_secret_uuid: "{{ cinder_ceph_client_uuid }}"

The following example sets cinder to use the cinder_volumes pool. The example uses cephx authentication and requires existing cinder account for cinder_volumes pool.

In user_variables.yml:

ceph_mons:
  - 172.29.244.151
  - 172.29.244.152
  - 172.29.244.153

In openstack_user_config.yml:

storage_hosts:
 infra1:
  ip: 172.29.236.101
  container_vars:
   cinder_backends:
     limit_container_types: cinder_volume
     rbd:
       volume_group: cinder-volumes
       volume_driver: cinder.volume.drivers.rbd.RBDDriver
       volume_backend_name: rbd
       rbd_pool: cinder-volumes
       rbd_ceph_conf: /etc/ceph/ceph.conf
       rbd_user: cinder
 infra2:
  ip: 172.29.236.102
  container_vars:
   cinder_backends:
     limit_container_types: cinder_volume
     rbd:
       volume_group: cinder-volumes
       volume_driver: cinder.volume.drivers.rbd.RBDDriver
       volume_backend_name: rbd
       rbd_pool: cinder-volumes
       rbd_ceph_conf: /etc/ceph/ceph.conf
       rbd_user: cinder
 infra3:
  ip: 172.29.236.103
  container_vars:
   cinder_backends:
     limit_container_types: cinder_volume
     rbd:
       volume_group: cinder-volumes
       volume_driver: cinder.volume.drivers.rbd.RBDDriver
       volume_backend_name: rbd
       rbd_pool: cinder-volumes
       rbd_ceph_conf: /etc/ceph/ceph.conf
       rbd_user: cinder

This link provides a complete working example of Ceph setup and integration with cinder (nova and glance included):

Configuring cinder to use Dell EqualLogic

To use the Dell EqualLogic volume driver as a back end, edit the /etc/openstack_deploy/openstack_user_config.yml file and configure the storage nodes that will use it.

Define the following parameters.

  1. Add dellqlx stanza under the cinder_backends for each storage node:

    cinder_backends:
  delleqlx:

  2. Specify volume back end name:

    volume_backend_name: DellEQLX_iSCSI

  3. Use Dell EQLX San ISCSI driver:

    volume_driver: cinder.volume.drivers.dell_emc.ps.PSSeriesISCSIDriver

  4. Specify the SAN IP address:

    san_ip: ip_of_dell_storage

  5. Specify SAN username (Default: grpadmin):

    san_login: grpadmin

  6. Specify the SAN password:

    san_password: password

  7. Specify the group name for pools (Default: group-0):

    eqlx_group_name: group-0

  8. Specify the pool where Cinder will create volumes and snapshots (Default: default):

    eqlx_pool: default

  9. Ensure the openstack_user_config.yml configuration is accurate:

    storage_hosts:
  Infra01:
    ip: infra_host_ip
    container_vars:
      cinder_backends:
        limit_container_types: cinder_volume
        delleqlx:
          volume_backend_name: DellEQLX_iSCSI
          volume_driver: cinder.volume.drivers.dell_emc.ps.PSSeriesISCSIDriver
          san_ip: ip_of_dell_storage
          san_login: grpadmin
          san_password: password
          eqlx_group_name: group-0
          eqlx_pool: default

Note

For more details about available configuration options, see https://docs.openstack.org/ocata/config-reference/block-storage/drivers/dell-equallogic-driver.html

Configuring cinder to use a NetApp appliance

To use a NetApp storage appliance back end, edit the /etc/openstack_deploy/openstack_user_config.yml file and configure each storage node that will use it.

Note

Ensure that the NAS Team enables httpd.admin.access.

  1. Add the netapp stanza under the cinder_backends stanza for each storage node:

    cinder_backends:
  netapp:

    The options in subsequent steps fit under the netapp stanza.

    The backend name is arbitrary and becomes a volume type within cinder.

  2. Configure the storage family:

    netapp_storage_family: STORAGE_FAMILY

    Replace STORAGE_FAMILY with ontap_7mode for Data ONTAP operating in 7-mode or ontap_cluster for Data ONTAP operating as a cluster.

  3. Configure the storage protocol:

    netapp_storage_protocol: STORAGE_PROTOCOL

    Replace STORAGE_PROTOCOL with iscsi for iSCSI or nfs for NFS.

    For the NFS protocol, specify the location of the configuration file that lists the shares available to cinder:

    nfs_shares_config: FILENAME_NFS_SHARES

    Replace FILENAME_NFS_SHARES with the location of the share configuration file. For example, /etc/cinder/nfs_shares.

  4. Configure the server:

    netapp_server_hostname: SERVER_HOSTNAME

    Replace SERVER_HOSTNAME with the hostnames for both netapp controllers.

  5. Configure the server API port:

    netapp_server_port: PORT_NUMBER

    Replace PORT_NUMBER with 80 for HTTP or 443 for HTTPS.

  6. Configure the server credentials:

    netapp_login: USER_NAME
netapp_password: PASSWORD

    Replace USER_NAME and PASSWORD with the appropriate values.

  7. Select the NetApp driver:

    volume_driver: cinder.volume.drivers.netapp.common.NetAppDriver

  8. Configure the volume back end name:

    volume_backend_name: BACKEND_NAME

    Replace BACKEND_NAME with a value that provides a hint for the cinder scheduler. For example, NETAPP_iSCSI.

  9. Ensure the openstack_user_config.yml configuration is accurate:

    storage_hosts:
  Infra01:
    ip: 172.29.236.16
    container_vars:
      cinder_backends:
        limit_container_types: cinder_volume
        netapp:
          netapp_storage_family: ontap_7mode
          netapp_storage_protocol: nfs
          netapp_server_hostname: 111.222.333.444
          netapp_server_port: 80
          netapp_login: openstack_cinder
          netapp_password: password
          volume_driver: cinder.volume.drivers.netapp.common.NetAppDriver
          volume_backend_name: NETAPP_NFS

    For netapp_server_hostname, specify the IP address of the Data ONTAP server. Include iSCSI or NFS for the netapp_storage_family depending on the configuration. Add 80 if using HTTP or 443 if using HTTPS for netapp_server_port.

    The cinder-volume.yml playbook will automatically install the nfs-common file across the hosts, transitioning from an LVM to a NetApp back end.

Configuring cinder qos specs

Deployers may optionally define the variable cinder_qos_specs to create qos specs. This variable is a list of dictionaries that contain the options for each qos spec. cinder volume-types may be assigned to a qos spec by defining the key cinder_volume_types in the desired qos spec dictionary.

- name: high-iops
  options:
    consumer: front-end
    read_iops_sec: 2000
    write_iops_sec: 2000
  cinder_volume_types:
    - volumes-1
    - volumes-2
- name: low-iops
  options:
    consumer: front-end
    write_iops_sec: 100

Shared storage and synchronized UID/GID

Specify a custom UID for the cinder user and GID for the cinder group to ensure they are identical on each host. This is helpful when using shared storage on Compute nodes because it allows instances to migrate without filesystem ownership failures.

By default, Ansible creates the cidner user and group without specifying the UID or GID. To specify custom values for the UID or GID, set the following Ansible variables:

cinder_system_user_uid = <specify a UID>
cinder_system_group_gid = <specify a GID>

Warning

Setting this value after deploying an environment with OpenStack-Ansible can cause failures, errors, and general instability. These values should only be set once before deploying an OpenStack environment and then never changed.

View Git Blame Copy Permalink