Nimble Storage fully integrates with the OpenStack platform through the Nimble Cinder driver, allowing a host to configure and manage Nimble and Alletra 6k Storage array features through Block Storage interfaces.
Support for iSCSI storage protocol is available with NimbleISCSIDriver Volume Driver class and Fibre Channel with NimbleFCDriver.
Support for the Liberty release and above is available from Nimble OS 2.3.8 or later.
Support for the Ocata release and above is available from Nimble OS 3.6 or later.
For Xena release, Nimble OS 5.3 or later is used and Alletra OS 6.0 or later is used.
Nimble and Alletra 6k Storage Cinder driver does not support port binding with multiple interfaces on the same subnet due to existing limitation in os-brick. This is partially referenced in the bug https://bugs.launchpad.net/os-brick/+bug/1722432 but does not resolve for multiple software iscsi ifaces.
Create, delete, clone, attach, and detach volumes
Create and delete volume snapshots
Create a volume from a snapshot
Copy an image to a volume
Copy a volume to an image
Extend a volume
Get volume statistics
Manage and unmanage a volume
Enable encryption and default performance policy for a volume-type extra-specs
Force backup of an in-use volume
Retype a volume
Create a Thinly Provisioned Volume
Attach a volume to multiple servers simultaneously (multiattach)
Volume Revert to Snapshot
Create, list, update, and delete consistency groups
Create, list, and delete consistency group snapshots
Update the file /etc/cinder/cinder.conf
with the given configuration.
Note: These parameters apply to Alletra 6k Storage as well.
In case of a basic (single back-end) configuration, add the parameters
within the [default]
section as follows.
[default]
san_ip = NIMBLE_MGMT_IP
san_login = NIMBLE_USER
san_password = NIMBLE_PASSWORD
use_multipath_for_image_xfer = True
volume_driver = NIMBLE_VOLUME_DRIVER
san_thin_provision = True
In case of multiple back-end configuration, for example, configuration which supports multiple Nimble Storage arrays or a single Nimble Storage array with arrays from other vendors, use the following parameters.
[default]
enabled_backends = Nimble-Cinder
[Nimble-Cinder]
san_ip = NIMBLE_MGMT_IP
san_login = NIMBLE_USER
san_password = NIMBLE_PASSWORD
use_multipath_for_image_xfer = True
volume_driver = NIMBLE_VOLUME_DRIVER
volume_backend_name = NIMBLE_BACKEND_NAME
In case of multiple back-end configuration, Nimble Storage volume type is created and associated with a back-end name as follows.
Note
Single back-end configuration users do not need to create the volume type.
$ openstack volume type create NIMBLE_VOLUME_TYPE
$ openstack volume type set --property volume_backend_name=NIMBLE_BACKEND_NAME NIMBLE_VOLUME_TYPE
This section explains the variables used above:
Management IP address of Nimble/Alletra 6k Storage array/group.
Nimble/Alletra 6k Storage account login with minimum power user
(admin) privilege if RBAC is used.
Password of the admin account for Nimble/Alletra 6k array.
Use either cinder.volume.drivers.hpe.nimble.NimbleISCSIDriver for iSCSI or cinder.volume.drivers.hpe.nimble.NimbleFCDriver for Fibre Channel.
A volume back-end name which is specified in the cinder.conf
file.
This is also used while assigning a back-end name to the Nimble volume type.
The Nimble volume-type which is created from the CLI and associated with
NIMBLE_BACKEND_NAME
.
Note
Restart the cinder-api
, cinder-scheduler
, and cinder-volume
services after updating the cinder.conf
file.
The Nimble volume driver also supports the following extra spec options:
Used to enable encryption for a volume-type.
PERF_POL_NAME is the name of a performance policy which exists on the Nimble/Alletra 6k array and should be enabled for every volume in a volume type.
Note
When upgrading to OpenStack deployment to Victoria or later,
do unset nimble:multi-initiator
extra-spec and set multiattach='<is> True'
.
Used to enable dedupe support for a volume-type.
Used to set the IOPS_LIMIT between 256 and 4294967294 for all volumes created for this volume-type.
FOLDER_NAME is the name of the folder which exists on the Nimble/Alletra 6k array and should be enabled for every volume in a volume type
These extra-specs can be enabled by using the following command:
$ openstack volume type set --property KEY=VALUE VOLUME_TYPE
VOLUME_TYPE
is the Nimble volume type and KEY
and VALUE
are
the options mentioned above.
The Nimble/Alletra 6k storage driver supports these configuration options:
Configuration option = Default value |
Description |
---|---|
|
(String) Nimble Controller pool name |
|
(String) Nimble Subnet Label |
|
(String) Path to Nimble Array SSL certificate |
|
(Boolean) Whether to verify Nimble SSL Certificate |
In OpenStack environments where Cinder block device multipathing is desired there are a few things to consider.
Configuring mulitpathing varies by system depending on the environment. In a
scenario where solely Nimble devices are being created by Cinder, the
following /etc/multipath.conf
file may be used:
defaults {
user_friendly_names yes
find_multipaths no
}
blacklist {
devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st)[0-9]*"
devnode "^hd[a-z]"
device {
vendor ".*"
product ".*"
}
}
blacklist_exceptions {
device {
vendor "Nimble"
product "Server"
}
}
devices {
device {
vendor "Nimble"
product "Server"
path_grouping_policy group_by_prio
prio "alua"
hardware_handler "1 alua"
path_selector "service-time 0"
path_checker tur
features "1 queue_if_no_path"
no_path_retry 30
failback immediate
fast_io_fail_tmo 5
dev_loss_tmo infinity
rr_min_io_rq 1
rr_weight uniform
}
}
After making changes to /etc/multipath.conf
, the multipath subsystem needs
to be reconfigured:
# multipathd reconfigure
Tip
The latest best practices for Nimble devices can be found in the HPE Nimble Storage Linux Integration Guide found on https://infosight.hpe.com
Important
OpenStack Cinder is currently not compatible with the HPE Nimble Storage Linux Toolkit (NLT)
Nova needs to be configured to pickup the actual multipath device created on the host.
In /etc/nova/nova.conf
, add the following to the [libvirt]
section:
[libvirt]
volume_use_multipath = True
Note
In versions prior to Newton, the option was called iscsi_use_multipath
After editing the Nova configuration file, the nova-conductor
service
needs to be restarted.
Tip
Depending on which particular OpenStack distribution is being used, Nova may use a different configuration file than the default.
To validate that instances get properly connected to the multipath device, inspect the instance devices:
# virsh dumpxml <Instance ID | Instance Name | Instance UUID>
Except where otherwise noted, this document is licensed under Creative Commons Attribution 3.0 License. See all OpenStack Legal Documents.