df686019f1
For the install documents, all the 'ironic' CLI commands are replaced with their equivalent 'openstack baremetal' CLI commands. Change-Id: Ic49cfd8d94366acb36c9a20b5c76885d7440f8d5 Partial-Bug: #1711235
31 KiB
Enrollment
After all the services have been properly configured, you should
enroll your hardware with the Bare Metal service, and confirm that the
Compute service sees the available hardware. The nodes will be visible
to the Compute service once they are in the available
provision state.
Note
After enrolling nodes with the Bare Metal service, the Compute
service will not be immediately notified of the new resources. The
Compute service's resource tracker syncs periodically, and so any
changes made directly to the Bare Metal service's resources will become
visible in the Compute service only after the next run of that periodic
task. More information is in the troubleshooting-install section.
Note
Any bare metal node that is visible to the Compute service may have a
workload scheduled to it, if both the power and
deploy interfaces pass the validate check. If
you wish to exclude a node from the Compute service's scheduler, for
instance so that you can perform maintenance on it, you can set the node
to "maintenance" mode. For more information see the maintenance_mode section.
Choosing a driver
When enrolling a node, the most important information to supply is
driver. This can be either a classic driver or a
hardware type - see enabling-drivers for the difference. The
driver list command can be used to list all drivers (of
both types) enabled on all hosts:
openstack baremetal driver list
+---------------------+-----------------------+
| Supported driver(s) | Active host(s) |
+---------------------+-----------------------+
| ipmi | localhost.localdomain |
| pxe_ipmitool | localhost.localdomain |
+---------------------+-----------------------+Starting with API version 1.31 (and python-ironicclient
1.13), you can also list only classic or only dynamic drivers:
openstack baremetal --os-baremetal-api-version 1.31 driver list --type dynamic
+---------------------+-----------------------+
| Supported driver(s) | Active host(s) |
+---------------------+-----------------------+
| ipmi | localhost.localdomain |
+---------------------+-----------------------+The specific driver to use should be picked based on actual hardware
capabilities and expected features. See /admin/drivers for more hints on that.
Each driver has a list of driver properties that need to be
specified via the node's driver_info field, in order for
the driver to operate on node. This list consists of the properties of
the hardware interfaces that the driver uses. These driver properties
are available with the driver property list command:
$ openstack baremetal driver property list pxe_ipmitool
+----------------------+-------------------------------------------------------------------------------------------------------------+
| Property | Description |
+----------------------+-------------------------------------------------------------------------------------------------------------+
| ipmi_address | IP address or hostname of the node. Required. |
| ipmi_password | password. Optional. |
| ipmi_username | username; default is NULL user. Optional. |
| ... | ... |
| deploy_kernel | UUID (from Glance) of the deployment kernel. Required. |
| deploy_ramdisk | UUID (from Glance) of the ramdisk that is mounted at boot time. Required. |
+----------------------+-------------------------------------------------------------------------------------------------------------+The properties marked as required must be supplied either during node creation or shortly after. Some properties may only be required for certain features.
Note on API versions
Starting with API version 1.11, the Bare Metal service added a new
initial provision state of enroll to its state machine.
When this or later API version is used, new nodes get this state instead
of available.
Existing automation tooling that use an API version lower than 1.11
are not affected, since the initial provision state is still
available. However, using API version 1.11 or above may
break existing automation tooling with respect to node creation.
The default API version used by (the most recent) python-ironicclient is 1.9, but it may change in the future and should not be relied on.
In the examples below we will use version 1.11 of the Bare metal API. This gives us the following advantages:
- Explicit power credentials validation before leaving the
enrollstate. - Running node cleaning before entering the
availablestate. - Not exposing half-configured nodes to the scheduler.
To set the API version for all commands, you can set the environment
variable IRONIC_API_VERSION. For the OpenStackClient
baremetal plugin, set the OS_BAREMETAL_API_VERSION variable
to the same value. For example:
$ export IRONIC_API_VERSION=1.11
$ export OS_BAREMETAL_API_VERSION=1.11Enrollment process
Creating a node
This section describes the main steps to enroll a node and make it available for provisioning. Some steps are shown separately for illustration purposes, and may be combined if desired.
Create a node in the Bare Metal service with the
node createcommand. At a minimum, you must specify the driver name (for example,pxe_ipmitool,agent_ipmitooloripmi).This command returns the node UUID along with other information about the node. The node's provision state will be
enroll:$ export OS_BAREMETAL_API_VERSION=1.11 $ openstack baremetal node create --driver pxe_ipmitool +--------------+--------------------------------------+ | Property | Value | +--------------+--------------------------------------+ | uuid | dfc6189f-ad83-4261-9bda-b27258eb1987 | | driver_info | {} | | extra | {} | | driver | pxe_ipmitool | | chassis_uuid | | | properties | {} | | name | None | +--------------+--------------------------------------+ $ openstack baremetal node show dfc6189f-ad83-4261-9bda-b27258eb1987 +------------------------+--------------------------------------+ | Property | Value | +------------------------+--------------------------------------+ | target_power_state | None | | extra | {} | | last_error | None | | maintenance_reason | None | | provision_state | enroll | | uuid | dfc6189f-ad83-4261-9bda-b27258eb1987 | | console_enabled | False | | target_provision_state | None | | provision_updated_at | None | | maintenance | False | | power_state | None | | driver | pxe_ipmitool | | properties | {} | | instance_uuid | None | | name | None | | driver_info | {} | | ... | ... | +------------------------+--------------------------------------+A node may also be referred to by a logical name as well as its UUID. A name can be assigned to the node during its creation by adding the
-noption to thenode createcommand or by updating an existing node with thenode setcommand. See Logical Names for examples.Starting with API version 1.31 (and
python-ironicclient1.13), you can pick which hardware interface to use with nodes that use hardware types. Each interface is represented by a node field called<IFACE>_interfacewhere<IFACE>in the interface type, e.g.boot. Seeenabling-driversfor details on hardware interfaces and/admin/upgrade-to-hardware-typesfor the matching between classic drivers and hardware types.An interface can be set either separately:
$ openstack baremetal --os-baremetal-api-version 1.31 node set $NODE_UUID \ --deploy-interface direct \ --raid-interface agentor set during node creation:
$ openstack baremetal --os-baremetal-api-version 1.31 node create --driver ipmi \ --deploy-interface direct \ --raid-interface agentIf no value is provided for some interfaces, Defaults for hardware interfaces are used instead.
It's an error to try changing this field for a node with a classic driver, and setting node's driver to classic one causes these fields to be set to
Noneautomatically.Update the node
driver_infowith the required driver properties, so that the Bare Metal service can manage the node:$ openstack baremetal node set $NODE_UUID \ --driver-info ipmi_username=$USER \ --driver-info ipmi_password=$PASS \ --driver-info ipmi_address=$ADDRESSNote
If IPMI is running on a port other than 623 (the default). The port must be added to
driver_infoby specifying theipmi_portvalue. Example:$ openstack baremetal node set $NODE_UUID --driver-info ipmi_port=$PORT_NUMBERYou may also specify all
driver_infoparameters during node creation by passing the --driver-info option multiple times:$ openstack baremetal node create --driver pxe_ipmitool \ --driver-info ipmi_username=$USER \ --driver-info ipmi_password=$PASS \ --driver-info ipmi_address=$ADDRESSSee Choosing a driver above for details on driver properties.
Specify a deploy kernel and ramdisk compatible with the node's driver, for example:
$ openstack baremetal node set $NODE_UUID \ --driver-info deploy_kernel=$DEPLOY_VMLINUZ_UUID \ --driver-info deploy_ramdisk=$DEPLOY_INITRD_UUIDSee
configure-glance-imagesfor details.You must also inform the Bare Metal service of the network interface cards which are part of the node by creating a port with each NIC's MAC address. These MAC addresses are passed to the Networking service during instance provisioning and used to configure the network appropriately:
$ openstack baremetal port create $MAC_ADDRESS --node $NODE_UUID
Adding scheduling information
Assign a resource class to the node. A resource class should represent a class of hardware in your data center, that corresponds to a Compute flavor.
For example, let's split hardware into these three groups:
- nodes with a lot of RAM and powerful CPU for computational tasks,
- nodes with powerful GPU for OpenCL computing,
- smaller nodes for development and testing.
We can define three resource classes to reflect these hardware groups, named
large-cpu,large-gpuandsmallrespectively. Then, for each node in each of the hardware groups, we'll set theirresource_classappropriately via:$ openstack --os-baremetal-api-version 1.21 baremetal node set $NODE_UUID \ --resource-class $CLASS_NAMEThe
--resource-classargument can also be used when creating a node:$ openstack --os-baremetal-api-version 1.21 baremetal node create \ --driver $DRIVER --resource-class $CLASS_NAMETo use resource classes for scheduling you need to update your flavors as described in
configure-nova-flavors.Warning
Scheduling based on resource classes will replace scheduling based on properties in the Queens release.
Note
This is not required for standalone deployments, only for those using the Compute service for provisioning bare metal instances.
Update the node's properties to match the actual hardware of the node:
$ openstack baremetal node set $NODE_UUID \ --property cpus=$CPU_COUNT \ --property memory_mb=$RAM_MB \ --property local_gb=$DISK_GB \ --property cpu_arch=$ARCHAs above, these can also be specified at node creation by passing the --property option to
node createmultiple times:$ openstack baremetal node create --driver pxe_ipmitool \ --driver-info ipmi_username=$USER \ --driver-info ipmi_password=$PASS \ --driver-info ipmi_address=$ADDRESS \ --property cpus=$CPU_COUNT \ --property memory_mb=$RAM_MB \ --property local_gb=$DISK_GB \ --property cpu_arch=$ARCHThese values can also be discovered during Hardware Inspection.
Warning
If scheduling based on resource classes is not used, the three properties
cpus,memory_mbandlocal_gbmust match ones defined on the flavor created whenconfigure-nova-flavors.Warning
The value provided for the
local_gbproperty must match the size of the root device you're going to deploy on. By default ironic-python-agent picks the smallest disk which is not smaller than 4 GiB.If you override this logic by using root device hints (see
root-device-hints), thelocal_gbvalue should match the size of picked target disk.If you wish to perform more advanced scheduling of the instances based on hardware capabilities, you may add metadata to each node that will be exposed to the the Compute scheduler (see: ComputeCapabilitiesFilter). A full explanation of this is outside of the scope of this document. It can be done through the special
capabilitiesmember of node properties:$ openstack baremetal node set $NODE_UUID \ --property capabilities=key1:val1,key2:val2Some capabilities can also be discovered during Hardware Inspection.
Validating node information
To check if Bare Metal service has the minimum information necessary for a node's driver to be functional, you may
validateit:$ openstack baremetal node validate $NODE_UUID +------------+--------+--------+ | Interface | Result | Reason | +------------+--------+--------+ | console | True | | | deploy | True | | | management | True | | | power | True | | +------------+--------+--------+If the node fails validation, each driver interface will return information as to why it failed:
$ openstack baremetal node validate $NODE_UUID +------------+--------+-------------------------------------------------------------------------------------------------------------------------------------+ | Interface | Result | Reason | +------------+--------+-------------------------------------------------------------------------------------------------------------------------------------+ | console | None | not supported | | deploy | False | Cannot validate iSCSI deploy. Some parameters were missing in node's instance_info. Missing are: ['root_gb', 'image_source'] | | management | False | Missing the following IPMI credentials in node's driver_info: ['ipmi_address']. | | power | False | Missing the following IPMI credentials in node's driver_info: ['ipmi_address']. | +------------+--------+-------------------------------------------------------------------------------------------------------------------------------------+When using the Compute Service with the Bare Metal service, it is safe to ignore the deploy interface's validation error due to lack of image information. You may continue the enrollment process. This information will be set by the Compute Service just before deploying, when an instance is requested:
$ openstack baremetal node validate $NODE_UUID +------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | Interface | Result | Reason | +------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | console | True | | | deploy | False | Cannot validate image information for node because one or more parameters are missing from its instance_info. Missing are: ['ramdisk', 'kernel', 'image_source'] | | management | True | | | power | True | | +------------+--------+------------------------------------------------------------------------------------------------------------------------------------------------------------------+
Making node available for deployment
In order for nodes to be available for deploying workloads on them,
nodes must be in the available provision state. To do this,
nodes created with API version 1.11 and above must be moved from the
enroll state to the manageable state and then
to the available state. This section can be safely skipped,
if API version 1.10 or earlier is used (which is the case by
default).
After creating a node and before moving it from its initial provision
state of enroll, basic power and port information needs to
be configured on the node. The Bare Metal service needs this information
because it verifies that it is capable of controlling the node when
transitioning the node from enroll to
manageable state.
To move a node from enroll to manageable
provision state:
$ openstack baremetal --os-baremetal-api-version 1.11 node manage $NODE_UUID
$ openstack baremetal node show $NODE_UUID
+------------------------+--------------------------------------------------------------------+
| Property | Value |
+------------------------+--------------------------------------------------------------------+
| ... | ... |
| provision_state | manageable | <- verify correct state
| uuid | 0eb013bb-1e4b-4f4c-94b5-2e7468242611 |
| ... | ... |
+------------------------+--------------------------------------------------------------------+Note
Since it is an asynchronous call, the response for
openstack baremetal node manage will not indicate whether
the transition succeeded or not. You can check the status of the
operation via openstack baremetal node show. If it was
successful, provision_state will be in the desired state.
If it failed, there will be information in the node's
last_error.
When a node is moved from the manageable to
available provision state, the node will go through
automated cleaning if configured to do so (see configure-cleaning).
To move a node from manageable to available
provision state:
$ openstack baremetal --os-baremetal-api-version 1.11 node provide $NODE_UUID
$ openstack baremetal node show $NODE_UUID
+------------------------+--------------------------------------------------------------------+
| Property | Value |
+------------------------+--------------------------------------------------------------------+
| ... | ... |
| provision_state | available | < - verify correct state
| uuid | 0eb013bb-1e4b-4f4c-94b5-2e7468242611 |
| ... | ... |
+------------------------+--------------------------------------------------------------------+For more details on the Bare Metal service's state machine, see the
/contributor/states
documentation.
Logical names
A node may also be referred to by a logical name as well as its UUID.
Names can be assigned either during its creation by adding the
-n option to the node create command or by
updating an existing node with the node set command.
Node names must be unique, and conform to:
The node is named 'example' in the following examples:
$ openstack baremetal node create --driver agent_ipmitool --name exampleor
$ openstack baremetal node set $NODE_UUID --name exampleOnce assigned a logical name, a node can then be referred to by name or UUID interchangeably:
$ openstack baremetal node create --driver agent_ipmitool --name example
+--------------+--------------------------------------+
| Property | Value |
+--------------+--------------------------------------+
| uuid | 71e01002-8662-434d-aafd-f068f69bb85e |
| driver_info | {} |
| extra | {} |
| driver | agent_ipmitool |
| chassis_uuid | |
| properties | {} |
| name | example |
+--------------+--------------------------------------+
$ openstack baremetal node show example
+------------------------+--------------------------------------+
| Property | Value |
+------------------------+--------------------------------------+
| target_power_state | None |
| extra | {} |
| last_error | None |
| updated_at | 2015-04-24T16:23:46+00:00 |
| ... | ... |
| instance_info | {} |
+------------------------+--------------------------------------+Defaults for hardware interfaces
For classic drivers all hardware interface implementations (except for the network interface) are hardcoded and cannot be changed. For hardware types, users can request one of enabled implementations when creating or updating a node as explained in Creating a node.
When no value is provided for a certain interface when creating a node, or changing a node's hardware type, the default value is used. You can use the driver details command to list the current enabled and default interfaces for a hardware type (for your deployment):
$ openstack baremetal --os-baremetal-api-version 1.31 driver show ipmi
+-------------------------------+----------------+
| Field | Value |
+-------------------------------+----------------+
| default_boot_interface | pxe |
| default_console_interface | no-console |
| default_deploy_interface | iscsi |
| default_inspect_interface | no-inspect |
| default_management_interface | ipmitool |
| default_network_interface | flat |
| default_power_interface | ipmitool |
| default_raid_interface | no-raid |
| default_vendor_interface | no-vendor |
| enabled_boot_interfaces | pxe |
| enabled_console_interfaces | no-console |
| enabled_deploy_interfaces | iscsi, direct |
| enabled_inspect_interfaces | no-inspect |
| enabled_management_interfaces | ipmitool |
| enabled_network_interfaces | flat, noop |
| enabled_power_interfaces | ipmitool |
| enabled_raid_interfaces | no-raid, agent |
| enabled_vendor_interfaces | no-vendor |
| hosts | ironic-host-1 |
| name | ipmi |
| type | dynamic |
+-------------------------------+----------------+The defaults are calculated as follows:
If the
default_<IFACE>_interfaceconfiguration option (where<IFACE>is the interface name) is set, its value is used as the default.If this implementation is not compatible with the node's hardware type, an error is returned to a user. An explicit value has to be provided for the node's
<IFACE>_interfacefield in this case.Otherwise, the first supported implementation that is enabled by an operator is used as the default.
A list of supported implementations is calculated by taking the intersection between the implementations supported by the node's hardware type and implementations enabled by the
enabled_<IFACE>_interfacesoption (where<IFACE>is the interface name). The calculation preserves the order of items, as provided by the hardware type.If the list of supported implementations is not empty, the first one is used. Otherwise, an error is returned to a user. In this case, an explicit value has to be provided for the
<IFACE>_interfacefield.
See enabling-drivers
for more details on configuration.
Example
Consider the following configuration (shortened for simplicity):
[DEFAULT]
enabled_hardware_types = ipmi,redfish
enabled_console_interfaces = no-console,ipmitool-shellinabox
enabled_deploy_interfaces = iscsi,direct
enabled_management_interfaces = ipmitool,redfish
enabled_power_interfaces = ipmitool,redfish
default_deploy_interface = directA new node is created with the ipmi driver and no
interfaces specified:
$ export OS_BAREMETAL_API_VERSION=1.31
$ openstack baremetal node create --driver ipmi
+--------------+--------------------------------------+
| Property | Value |
+--------------+--------------------------------------+
| uuid | dfc6189f-ad83-4261-9bda-b27258eb1987 |
| driver_info | {} |
| extra | {} |
| driver | ipmi |
| chassis_uuid | |
| properties | {} |
| name | None |
+--------------+--------------------------------------+Then the defaults for the interfaces that will be used by the node in this example are calculated as follows:
- deploy
-
An explicit value of
directis provided fordefault_deploy_interface, so it is used. - power
-
No default is configured. The
ipmihardware type supports onlyipmitoolpower. The intersection between supported power interfaces and values provided in theenabled_power_interfacesoption has only one item:ipmitool. It is used. - console
-
No default is configured. The
ipmihardware type supports the following console interfaces:ipmitool-socat,ipmitool-shellinaboxandno-console(in this order). Of these three, only two are enabled:no-consoleandipmitool-shellinabox(order does not matter). The intersection containsipmitool-shellinaboxandno-console. The first item is used, and it isipmitool-shellinabox. - management
-
Following the same calculation as power, the
ipmitoolmanagement interface is used.
Hardware Inspection
The Bare Metal service supports hardware inspection that simplifies
enrolling nodes - please see /admin/inspection for details.
Tenant Networks and Port Groups
See /admin/multitenancy and /admin/portgroups.