salt/states/esxi.py
# -*- coding: utf-8 -*-
'''
Manage VMware ESXi Hosts.
.. versionadded:: 2015.8.4
Dependencies
============
- pyVmomi Python Module
- ESXCLI
pyVmomi
-------
PyVmomi can be installed via pip:
.. code-block:: bash
pip install pyVmomi
.. note::
Version 6.0 of pyVmomi has some problems with SSL error handling on certain
versions of Python. If using version 6.0 of pyVmomi, Python 2.6,
Python 2.7.9, or newer must be present. This is due to an upstream dependency
in pyVmomi 6.0 that is not supported in Python versions 2.7 to 2.7.8. If the
version of Python is not in the supported range, you will need to install an
earlier version of pyVmomi. See `Issue #29537`_ for more information.
.. _Issue #29537: https://github.com/saltstack/salt/issues/29537
Based on the note above, to install an earlier version of pyVmomi than the
version currently listed in PyPi, run the following:
.. code-block:: bash
pip install pyVmomi==5.5.0.2014.1.1
The 5.5.0.2014.1.1 is a known stable version that this original ESXi State
Module was developed against.
ESXCLI
------
Currently, about a third of the functions used in the vSphere Execution Module require
the ESXCLI package be installed on the machine running the Proxy Minion process.
The ESXCLI package is also referred to as the VMware vSphere CLI, or vCLI. VMware
provides vCLI package installation instructions for `vSphere 5.5`_ and
`vSphere 6.0`_.
.. _vSphere 5.5: http://pubs.vmware.com/vsphere-55/index.jsp#com.vmware.vcli.getstart.doc/cli_install.4.2.html
.. _vSphere 6.0: http://pubs.vmware.com/vsphere-60/index.jsp#com.vmware.vcli.getstart.doc/cli_install.4.2.html
Once all of the required dependencies are in place and the vCLI package is
installed, you can check to see if you can connect to your ESXi host or vCenter
server by running the following command:
.. code-block:: bash
esxcli -s <host-location> -u <username> -p <password> system syslog config get
If the connection was successful, ESXCLI was successfully installed on your system.
You should see output related to the ESXi host's syslog configuration.
.. note::
Be aware that some functionality in this state module may depend on the
type of license attached to the ESXi host.
For example, certain services are only available to manipulate service state
or policies with a VMware vSphere Enterprise or Enterprise Plus license, while
others are available with a Standard license. The ``ntpd`` service is restricted
to an Enterprise Plus license, while ``ssh`` is available via the Standard
license.
Please see the `vSphere Comparison`_ page for more information.
.. _vSphere Comparison: https://www.vmware.com/products/vsphere/compare
About
-----
This state module was written to be used in conjunction with Salt's
:mod:`ESXi Proxy Minion <salt.proxy.esxi>`. For a tutorial on how to use Salt's
ESXi Proxy Minion, please refer to the
:ref:`ESXi Proxy Minion Tutorial <tutorial-esxi-proxy>` for
configuration examples, dependency installation instructions, how to run remote
execution functions against ESXi hosts via a Salt Proxy Minion, and a larger state
example.
'''
# Import Python Libs
from __future__ import absolute_import, print_function, unicode_literals
import logging
import sys
import re
# Import Salt Libs
from salt.ext import six
import salt.utils.files
from salt.exceptions import CommandExecutionError, InvalidConfigError, \
VMwareObjectRetrievalError, VMwareSaltError, VMwareApiError, \
ArgumentValueError
from salt.utils.decorators import depends
from salt.config.schemas.esxi import DiskGroupsDiskScsiAddressSchema, \
HostCacheSchema
# External libraries
try:
import jsonschema
HAS_JSONSCHEMA = True
except ImportError:
HAS_JSONSCHEMA = False
# Get Logging Started
log = logging.getLogger(__name__)
try:
from pyVmomi import VmomiSupport
# We check the supported vim versions to infer the pyVmomi version
if 'vim25/6.0' in VmomiSupport.versionMap and \
sys.version_info > (2, 7) and sys.version_info < (2, 7, 9):
log.debug('pyVmomi not loaded: Incompatible versions '
'of Python. See Issue #29537.')
raise ImportError()
HAS_PYVMOMI = True
except ImportError:
HAS_PYVMOMI = False
def __virtual__():
return 'esxi.cmd' in __salt__
def coredump_configured(name, enabled, dump_ip, host_vnic='vmk0', dump_port=6500):
'''
Ensures a host's core dump configuration.
name
Name of the state.
enabled
Sets whether or not ESXi core dump collection should be enabled.
This is a boolean value set to ``True`` or ``False`` to enable
or disable core dumps.
Note that ESXi requires that the core dump must be enabled before
any other parameters may be set. This also affects the ``changes``
results in the state return dictionary. If ``enabled`` is ``False``,
we can't obtain any previous settings to compare other state variables,
resulting in many ``old`` references returning ``None``.
Once ``enabled`` is ``True`` the ``changes`` dictionary comparisons
will be more accurate. This is due to the way the system coredemp
network configuration command returns data.
dump_ip
The IP address of host that will accept the dump.
host_vnic
Host VNic port through which to communicate. Defaults to ``vmk0``.
dump_port
TCP port to use for the dump. Defaults to ``6500``.
Example:
.. code-block:: yaml
configure-host-coredump:
esxi.coredump_configured:
- enabled: True
- dump_ip: 'my-coredump-ip.example.com'
'''
ret = {'name': name,
'result': False,
'changes': {},
'comment': ''}
esxi_cmd = 'esxi.cmd'
enabled_msg = 'ESXi requires that the core dump must be enabled ' \
'before any other parameters may be set.'
host = __pillar__['proxy']['host']
current_config = __salt__[esxi_cmd]('get_coredump_network_config').get(host)
error = current_config.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
current_config = current_config.get('Coredump Config')
current_enabled = current_config.get('enabled')
# Configure coredump enabled state, if there are changes.
if current_enabled != enabled:
enabled_changes = {'enabled': {'old': current_enabled, 'new': enabled}}
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('coredump_network_enable',
enabled=enabled).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
# Allow users to disable core dump, but then return since
# nothing else can be set if core dump is disabled.
if not enabled:
ret['result'] = True
ret['comment'] = enabled_msg
ret['changes'].update(enabled_changes)
return ret
ret['changes'].update(enabled_changes)
elif not enabled:
# If current_enabled and enabled match, but are both False,
# We must return before configuring anything. This isn't a
# failure as core dump may be disabled intentionally.
ret['result'] = True
ret['comment'] = enabled_msg
return ret
# Test for changes with all remaining configurations. The changes flag is used
# To detect changes, and then set_coredump_network_config is called one time.
changes = False
current_ip = current_config.get('ip')
if current_ip != dump_ip:
ret['changes'].update({'dump_ip':
{'old': current_ip,
'new': dump_ip}})
changes = True
current_vnic = current_config.get('host_vnic')
if current_vnic != host_vnic:
ret['changes'].update({'host_vnic':
{'old': current_vnic,
'new': host_vnic}})
changes = True
current_port = current_config.get('port')
if current_port != six.text_type(dump_port):
ret['changes'].update({'dump_port':
{'old': current_port,
'new': six.text_type(dump_port)}})
changes = True
# Only run the command if not using test=True and changes were detected.
if not __opts__['test'] and changes is True:
response = __salt__[esxi_cmd]('set_coredump_network_config',
dump_ip=dump_ip,
host_vnic=host_vnic,
dump_port=dump_port).get(host)
if response.get('success') is False:
msg = response.get('stderr')
if not msg:
msg = response.get('stdout')
ret['comment'] = 'Error: {0}'.format(msg)
return ret
ret['result'] = True
if ret['changes'] == {}:
ret['comment'] = 'Core Dump configuration is already in the desired state.'
return ret
if __opts__['test']:
ret['result'] = None
ret['comment'] = 'Core dump configuration will change.'
return ret
def password_present(name, password):
'''
Ensures the given password is set on the ESXi host. Passwords cannot be obtained from
host, so if a password is set in this state, the ``vsphere.update_host_password``
function will always run (except when using test=True functionality) and the state's
changes dictionary will always be populated.
The username for which the password will change is the same username that is used to
authenticate against the ESXi host via the Proxy Minion. For example, if the pillar
definition for the proxy username is defined as ``root``, then the username that the
password will be updated for via this state is ``root``.
name
Name of the state.
password
The new password to change on the host.
Example:
.. code-block:: yaml
configure-host-password:
esxi.password_present:
- password: 'new-bad-password'
'''
ret = {'name': name,
'result': True,
'changes': {'old': 'unknown',
'new': '********'},
'comment': 'Host password was updated.'}
esxi_cmd = 'esxi.cmd'
if __opts__['test']:
ret['result'] = None
ret['comment'] = 'Host password will change.'
return ret
else:
try:
__salt__[esxi_cmd]('update_host_password',
new_password=password)
except CommandExecutionError as err:
ret['result'] = False
ret['comment'] = 'Error: {0}'.format(err)
return ret
return ret
def ntp_configured(name,
service_running,
ntp_servers=None,
service_policy=None,
service_restart=False,
update_datetime=False):
'''
Ensures a host's NTP server configuration such as setting NTP servers, ensuring the
NTP daemon is running or stopped, or restarting the NTP daemon for the ESXi host.
name
Name of the state.
service_running
Ensures the running state of the ntp daemon for the host. Boolean value where
``True`` indicates that ntpd should be running and ``False`` indicates that it
should be stopped.
ntp_servers
A list of servers that should be added to the ESXi host's NTP configuration.
service_policy
The policy to set for the NTP service.
.. note::
When setting the service policy to ``off`` or ``on``, you *must* quote the
setting. If you don't, the yaml parser will set the string to a boolean,
which will cause trouble checking for stateful changes and will error when
trying to set the policy on the ESXi host.
service_restart
If set to ``True``, the ntp daemon will be restarted, regardless of its previous
running state. Default is ``False``.
update_datetime
If set to ``True``, the date/time on the given host will be updated to UTC.
Default setting is ``False``. This option should be used with caution since
network delays and execution delays can result in time skews.
Example:
.. code-block:: yaml
configure-host-ntp:
esxi.ntp_configured:
- service_running: True
- ntp_servers:
- 192.174.1.100
- 192.174.1.200
- service_policy: 'on'
- service_restart: True
'''
ret = {'name': name,
'result': False,
'changes': {},
'comment': ''}
esxi_cmd = 'esxi.cmd'
host = __pillar__['proxy']['host']
ntpd = 'ntpd'
ntp_config = __salt__[esxi_cmd]('get_ntp_config').get(host)
ntp_running = __salt__[esxi_cmd]('get_service_running',
service_name=ntpd).get(host)
error = ntp_running.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ntp_running = ntp_running.get(ntpd)
# Configure NTP Servers for the Host
if ntp_servers and set(ntp_servers) != set(ntp_config):
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('set_ntp_config',
ntp_servers=ntp_servers).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
# Set changes dictionary for ntp_servers
ret['changes'].update({'ntp_servers':
{'old': ntp_config,
'new': ntp_servers}})
# Configure service_running state
if service_running != ntp_running:
# Only run the command if not using test=True
if not __opts__['test']:
# Start ntdp if service_running=True
if ntp_running is True:
response = __salt__[esxi_cmd]('service_start',
service_name=ntpd).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
# Stop ntpd if service_running=False
else:
response = __salt__[esxi_cmd]('service_stop',
service_name=ntpd).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'service_running':
{'old': ntp_running,
'new': service_running}})
# Configure service_policy
if service_policy:
current_service_policy = __salt__[esxi_cmd]('get_service_policy',
service_name=ntpd).get(host)
error = current_service_policy.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
current_service_policy = current_service_policy.get(ntpd)
if service_policy != current_service_policy:
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('set_service_policy',
service_name=ntpd,
service_policy=service_policy).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'service_policy':
{'old': current_service_policy,
'new': service_policy}})
# Update datetime, if requested.
if update_datetime:
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('update_host_datetime').get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'update_datetime':
{'old': '',
'new': 'Host datetime was updated.'}})
# Restart ntp_service if service_restart=True
if service_restart:
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('service_restart',
service_name=ntpd).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'service_restart':
{'old': '',
'new': 'NTP Daemon Restarted.'}})
ret['result'] = True
if ret['changes'] == {}:
ret['comment'] = 'NTP is already in the desired state.'
return ret
if __opts__['test']:
ret['result'] = None
ret['comment'] = 'NTP state will change.'
return ret
def vmotion_configured(name, enabled, device='vmk0'):
'''
Configures a host's VMotion properties such as enabling VMotion and setting
the device VirtualNic that VMotion will use.
name
Name of the state.
enabled
Ensures whether or not VMotion should be enabled on a host as a boolean
value where ``True`` indicates that VMotion should be enabled and ``False``
indicates that VMotion should be disabled.
device
The device that uniquely identifies the VirtualNic that will be used for
VMotion for the host. Defaults to ``vmk0``.
Example:
.. code-block:: yaml
configure-vmotion:
esxi.vmotion_configured:
- enabled: True
- device: sample-device
'''
ret = {'name': name,
'result': False,
'changes': {},
'comment': ''}
esxi_cmd = 'esxi.cmd'
host = __pillar__['proxy']['host']
current_vmotion_enabled = __salt__[esxi_cmd]('get_vmotion_enabled').get(host)
current_vmotion_enabled = current_vmotion_enabled.get('VMotion Enabled')
# Configure VMotion Enabled state, if changed.
if enabled != current_vmotion_enabled:
# Only run the command if not using test=True
if not __opts__['test']:
# Enable VMotion if enabled=True
if enabled is True:
response = __salt__[esxi_cmd]('vmotion_enable',
device=device).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
# Disable VMotion if enabled=False
else:
response = __salt__[esxi_cmd]('vmotion_disable').get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'enabled':
{'old': current_vmotion_enabled,
'new': enabled}})
ret['result'] = True
if ret['changes'] == {}:
ret['comment'] = 'VMotion configuration is already in the desired state.'
return ret
if __opts__['test']:
ret['result'] = None
ret['comment'] = 'VMotion configuration will change.'
return ret
def vsan_configured(name, enabled, add_disks_to_vsan=False):
'''
Configures a host's VSAN properties such as enabling or disabling VSAN, or
adding VSAN-eligible disks to the VSAN system for the host.
name
Name of the state.
enabled
Ensures whether or not VSAN should be enabled on a host as a boolean
value where ``True`` indicates that VSAN should be enabled and ``False``
indicates that VSAN should be disabled.
add_disks_to_vsan
If set to ``True``, any VSAN-eligible disks for the given host will be added
to the host's VSAN system. Default is ``False``.
Example:
.. code-block:: yaml
configure-host-vsan:
esxi.vsan_configured:
- enabled: True
- add_disks_to_vsan: True
'''
ret = {'name': name,
'result': False,
'changes': {},
'comment': ''}
esxi_cmd = 'esxi.cmd'
host = __pillar__['proxy']['host']
current_vsan_enabled = __salt__[esxi_cmd]('get_vsan_enabled').get(host)
error = current_vsan_enabled.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
current_vsan_enabled = current_vsan_enabled.get('VSAN Enabled')
# Configure VSAN Enabled state, if changed.
if enabled != current_vsan_enabled:
# Only run the command if not using test=True
if not __opts__['test']:
# Enable VSAN if enabled=True
if enabled is True:
response = __salt__[esxi_cmd]('vsan_enable').get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
# Disable VSAN if enabled=False
else:
response = __salt__[esxi_cmd]('vsan_disable').get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'enabled':
{'old': current_vsan_enabled,
'new': enabled}})
# Add any eligible disks to VSAN, if requested.
if add_disks_to_vsan:
current_eligible_disks = __salt__[esxi_cmd]('get_vsan_eligible_disks').get(host)
error = current_eligible_disks.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
disks = current_eligible_disks.get('Eligible')
if disks and isinstance(disks, list):
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('vsan_add_disks').get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'add_disks_to_vsan':
{'old': '',
'new': disks}})
ret['result'] = True
if ret['changes'] == {}:
ret['comment'] = 'VSAN configuration is already in the desired state.'
return ret
if __opts__['test']:
ret['result'] = None
ret['comment'] = 'VSAN configuration will change.'
return ret
def ssh_configured(name,
service_running,
ssh_key=None,
ssh_key_file=None,
service_policy=None,
service_restart=False,
certificate_verify=False):
'''
Manage the SSH configuration for a host including whether or not SSH is running or
the presence of a given SSH key. Note: Only one ssh key can be uploaded for root.
Uploading a second key will replace any existing key.
name
Name of the state.
service_running
Ensures whether or not the SSH service should be running on a host. Represented
as a boolean value where ``True`` indicates that SSH should be running and
``False`` indicates that SSH should stopped.
In order to update SSH keys, the SSH service must be running.
ssh_key
Public SSH key to added to the authorized_keys file on the ESXi host. You can
use ``ssh_key`` or ``ssh_key_file``, but not both.
ssh_key_file
File containing the public SSH key to be added to the authorized_keys file on
the ESXi host. You can use ``ssh_key_file`` or ``ssh_key``, but not both.
service_policy
The policy to set for the NTP service.
.. note::
When setting the service policy to ``off`` or ``on``, you *must* quote the
setting. If you don't, the yaml parser will set the string to a boolean,
which will cause trouble checking for stateful changes and will error when
trying to set the policy on the ESXi host.
service_restart
If set to ``True``, the SSH service will be restarted, regardless of its
previous running state. Default is ``False``.
certificate_verify
If set to ``True``, the SSL connection must present a valid certificate.
Default is ``False``.
Example:
.. code-block:: yaml
configure-host-ssh:
esxi.ssh_configured:
- service_running: True
- ssh_key_file: /etc/salt/ssh_keys/my_key.pub
- service_policy: 'on'
- service_restart: True
- certificate_verify: True
'''
ret = {'name': name,
'result': False,
'changes': {},
'comment': ''}
esxi_cmd = 'esxi.cmd'
host = __pillar__['proxy']['host']
ssh = 'ssh'
ssh_running = __salt__[esxi_cmd]('get_service_running',
service_name=ssh).get(host)
error = ssh_running.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ssh_running = ssh_running.get(ssh)
# Configure SSH service_running state, if changed.
if service_running != ssh_running:
# Only actually run the command if not using test=True
if not __opts__['test']:
# Start SSH if service_running=True
if service_running is True:
enable = __salt__[esxi_cmd]('service_start',
service_name=ssh).get(host)
error = enable.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
# Disable SSH if service_running=False
else:
disable = __salt__[esxi_cmd]('service_stop',
service_name=ssh).get(host)
error = disable.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'service_running':
{'old': ssh_running,
'new': service_running}})
# If uploading an SSH key or SSH key file, see if there's a current
# SSH key and compare the current key to the key set in the state.
current_ssh_key, ssh_key_changed = None, False
if ssh_key or ssh_key_file:
current_ssh_key = __salt__[esxi_cmd]('get_ssh_key',
certificate_verify=certificate_verify)
error = current_ssh_key.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
current_ssh_key = current_ssh_key.get('key')
if current_ssh_key:
clean_current_key = _strip_key(current_ssh_key).split(' ')
if not ssh_key:
ssh_key = ''
# Open ssh key file and read in contents to create one key string
with salt.utils.files.fopen(ssh_key_file, 'r') as key_file:
for line in key_file:
if line.startswith('#'):
# Commented line
continue
ssh_key = ssh_key + line
clean_ssh_key = _strip_key(ssh_key).split(' ')
# Check that the first two list items of clean key lists are equal.
if clean_current_key[0] != clean_ssh_key[0] or clean_current_key[1] != clean_ssh_key[1]:
ssh_key_changed = True
else:
# If current_ssh_key is None, but we're setting a new key with
# either ssh_key or ssh_key_file, then we need to flag the change.
ssh_key_changed = True
# Upload SSH key, if changed.
if ssh_key_changed:
if not __opts__['test']:
# Upload key
response = __salt__[esxi_cmd]('upload_ssh_key',
ssh_key=ssh_key,
ssh_key_file=ssh_key_file,
certificate_verify=certificate_verify)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'SSH Key':
{'old': current_ssh_key,
'new': ssh_key if ssh_key else ssh_key_file}})
# Configure service_policy
if service_policy:
current_service_policy = __salt__[esxi_cmd]('get_service_policy',
service_name=ssh).get(host)
error = current_service_policy.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
current_service_policy = current_service_policy.get(ssh)
if service_policy != current_service_policy:
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('set_service_policy',
service_name=ssh,
service_policy=service_policy).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'service_policy':
{'old': current_service_policy,
'new': service_policy}})
# Restart ssh_service if service_restart=True
if service_restart:
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('service_restart',
service_name=ssh).get(host)
error = response.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
ret['changes'].update({'service_restart':
{'old': '',
'new': 'SSH service restarted.'}})
ret['result'] = True
if ret['changes'] == {}:
ret['comment'] = 'SSH service is already in the desired state.'
return ret
if __opts__['test']:
ret['result'] = None
ret['comment'] = 'SSH service state will change.'
return ret
def syslog_configured(name,
syslog_configs,
firewall=True,
reset_service=True,
reset_syslog_config=False,
reset_configs=None):
'''
Ensures the specified syslog configuration parameters. By default,
this state will reset the syslog service after any new or changed
parameters are set successfully.
name
Name of the state.
syslog_configs
Name of parameter to set (corresponds to the command line switch for
esxcli without the double dashes (--))
Valid syslog_config values are ``logdir``, ``loghost``, ``logdir-unique``,
``default-rotate``, ``default-size``, and ``default-timeout``.
Each syslog_config option also needs a configuration value to set.
For example, ``loghost`` requires URLs or IP addresses to use for
logging. Multiple log servers can be specified by listing them,
comma-separated, but without spaces before or after commas
(reference: https://blogs.vmware.com/vsphere/2012/04/configuring-multiple-syslog-servers-for-esxi-5.html)
firewall
Enable the firewall rule set for syslog. Defaults to ``True``.
reset_service
After a successful parameter set, reset the service. Defaults to ``True``.
reset_syslog_config
Resets the syslog service to it's default settings. Defaults to ``False``.
If set to ``True``, default settings defined by the list of syslog configs
in ``reset_configs`` will be reset before running any other syslog settings.
reset_configs
A comma-delimited list of parameters to reset. Only runs if
``reset_syslog_config`` is set to ``True``. If ``reset_syslog_config`` is set
to ``True``, but no syslog configs are listed in ``reset_configs``, then
``reset_configs`` will be set to ``all`` by default.
See ``syslog_configs`` parameter above for a list of valid options.
Example:
.. code-block:: yaml
configure-host-syslog:
esxi.syslog_configured:
- syslog_configs:
loghost: ssl://localhost:5432,tcp://10.1.0.1:1514
default-timeout: 120
- firewall: True
- reset_service: True
- reset_syslog_config: True
- reset_configs: loghost,default-timeout
'''
ret = {'name': name,
'result': False,
'changes': {},
'comment': ''}
esxi_cmd = 'esxi.cmd'
host = __pillar__['proxy']['host']
if reset_syslog_config:
if not reset_configs:
reset_configs = 'all'
# Only run the command if not using test=True
if not __opts__['test']:
reset = __salt__[esxi_cmd]('reset_syslog_config',
syslog_config=reset_configs).get(host)
for key, val in six.iteritems(reset):
if isinstance(val, bool):
continue
if not val.get('success'):
msg = val.get('message')
if not msg:
msg = 'There was an error resetting a syslog config \'{0}\'.' \
'Please check debug logs.'.format(val)
ret['comment'] = 'Error: {0}'.format(msg)
return ret
ret['changes'].update({'reset_syslog_config':
{'old': '',
'new': reset_configs}})
current_firewall = __salt__[esxi_cmd]('get_firewall_status').get(host)
error = current_firewall.get('Error')
if error:
ret['comment'] = 'Error: {0}'.format(error)
return ret
current_firewall = current_firewall.get('rulesets').get('syslog')
if current_firewall != firewall:
# Only run the command if not using test=True
if not __opts__['test']:
enabled = __salt__[esxi_cmd]('enable_firewall_ruleset',
ruleset_enable=firewall,
ruleset_name='syslog').get(host)
if enabled.get('retcode') != 0:
err = enabled.get('stderr')
out = enabled.get('stdout')
ret['comment'] = 'Error: {0}'.format(err if err else out)
return ret
ret['changes'].update({'firewall':
{'old': current_firewall,
'new': firewall}})
current_syslog_config = __salt__[esxi_cmd]('get_syslog_config').get(host)
for key, val in six.iteritems(syslog_configs):
# The output of get_syslog_config has different keys than the keys
# Used to set syslog_config values. We need to look them up first.
try:
lookup_key = _lookup_syslog_config(key)
except KeyError:
ret['comment'] = '\'{0}\' is not a valid config variable.'.format(key)
return ret
current_val = current_syslog_config[lookup_key]
if six.text_type(current_val) != six.text_type(val):
# Only run the command if not using test=True
if not __opts__['test']:
response = __salt__[esxi_cmd]('set_syslog_config',
syslog_config=key,
config_value=val,
firewall=firewall,
reset_service=reset_service).get(host)
success = response.get(key).get('success')
if not success:
msg = response.get(key).get('message')
if not msg:
msg = 'There was an error setting syslog config \'{0}\'. ' \
'Please check debug logs.'.format(key)
ret['comment'] = msg
return ret
if not ret['changes'].get('syslog_config'):
ret['changes'].update({'syslog_config': {}})
ret['changes']['syslog_config'].update({key:
{'old': current_val,
'new': val}})
ret['result'] = True
if ret['changes'] == {}:
ret['comment'] = 'Syslog is already in the desired state.'
return ret
if __opts__['test']:
ret['result'] = None
ret['comment'] = 'Syslog state will change.'
return ret
@depends(HAS_PYVMOMI)
@depends(HAS_JSONSCHEMA)
def diskgroups_configured(name, diskgroups, erase_disks=False):
'''
Configures the disk groups to use for vsan.
This function will do the following:
1. Check whether or not all disks in the diskgroup spec exist, and raises
and errors if they do not.
2. Create diskgroups with the correct disk configurations if diskgroup
(identified by the cache disk canonical name) doesn't exist
3. Adds extra capacity disks to the existing diskgroup
Example:
.. code:: python
{
'cache_scsi_addr': 'vmhba1:C0:T0:L0',
'capacity_scsi_addrs': [
'vmhba2:C0:T0:L0',
'vmhba3:C0:T0:L0',
'vmhba4:C0:T0:L0',
]
}
name
Mandatory state name
diskgroups
Disk group representation containing scsi disk addresses.
Scsi addresses are expected for disks in the diskgroup:
erase_disks
Specifies whether to erase all partitions on all disks member of the
disk group before the disk group is created. Default value is False.
'''
proxy_details = __salt__['esxi.get_details']()
hostname = proxy_details['host'] if not proxy_details.get('vcenter') \
else proxy_details['esxi_host']
log.info('Running state %s for host \'%s\'', name, hostname)
# Variable used to return the result of the invocation
ret = {'name': name,
'result': None,
'changes': {},
'comments': None}
# Signals if errors have been encountered
errors = False
# Signals if changes are required
changes = False
comments = []
diskgroup_changes = {}
si = None
try:
log.trace('Validating diskgroups_configured input')
schema = DiskGroupsDiskScsiAddressSchema.serialize()
try:
jsonschema.validate({'diskgroups': diskgroups,
'erase_disks': erase_disks}, schema)
except jsonschema.exceptions.ValidationError as exc:
raise InvalidConfigError(exc)
si = __salt__['vsphere.get_service_instance_via_proxy']()
host_disks = __salt__['vsphere.list_disks'](service_instance=si)
if not host_disks:
raise VMwareObjectRetrievalError(
'No disks retrieved from host \'{0}\''.format(hostname))
scsi_addr_to_disk_map = {d['scsi_address']: d for d in host_disks}
log.trace('scsi_addr_to_disk_map = %s', scsi_addr_to_disk_map)
existing_diskgroups = \
__salt__['vsphere.list_diskgroups'](service_instance=si)
cache_disk_to_existing_diskgroup_map = \
{dg['cache_disk']: dg for dg in existing_diskgroups}
except CommandExecutionError as err:
log.error('Error: %s', err)
if si:
__salt__['vsphere.disconnect'](si)
ret.update({
'result': False if not __opts__['test'] else None,
'comment': six.text_type(err)})
return ret
# Iterate through all of the disk groups
for idx, dg in enumerate(diskgroups):
# Check for cache disk
if not dg['cache_scsi_addr'] in scsi_addr_to_disk_map:
comments.append('No cache disk with scsi address \'{0}\' was '
'found.'.format(dg['cache_scsi_addr']))
log.error(comments[-1])
errors = True
continue
# Check for capacity disks
cache_disk_id = scsi_addr_to_disk_map[dg['cache_scsi_addr']]['id']
cache_disk_display = '{0} (id:{1})'.format(dg['cache_scsi_addr'],
cache_disk_id)
bad_scsi_addrs = []
capacity_disk_ids = []
capacity_disk_displays = []
for scsi_addr in dg['capacity_scsi_addrs']:
if scsi_addr not in scsi_addr_to_disk_map:
bad_scsi_addrs.append(scsi_addr)
continue
capacity_disk_ids.append(scsi_addr_to_disk_map[scsi_addr]['id'])
capacity_disk_displays.append(
'{0} (id:{1})'.format(scsi_addr, capacity_disk_ids[-1]))
if bad_scsi_addrs:
comments.append('Error in diskgroup #{0}: capacity disks with '
'scsi addresses {1} were not found.'
''.format(idx,
', '.join(['\'{0}\''.format(a)
for a in bad_scsi_addrs])))
log.error(comments[-1])
errors = True
continue
if not cache_disk_to_existing_diskgroup_map.get(cache_disk_id):
# A new diskgroup needs to be created
log.trace('erase_disks = %s', erase_disks)
if erase_disks:
if __opts__['test']:
comments.append('State {0} will '
'erase all disks of disk group #{1}; '
'cache disk: \'{2}\', '
'capacity disk(s): {3}.'
''.format(name, idx, cache_disk_display,
', '.join(
['\'{}\''.format(a) for a in
capacity_disk_displays])))
else:
# Erase disk group disks
for disk_id in [cache_disk_id] + capacity_disk_ids:
__salt__['vsphere.erase_disk_partitions'](
disk_id=disk_id, service_instance=si)
comments.append('Erased disks of diskgroup #{0}; '
'cache disk: \'{1}\', capacity disk(s): '
'{2}'.format(
idx, cache_disk_display,
', '.join(['\'{0}\''.format(a) for a in
capacity_disk_displays])))
log.info(comments[-1])
if __opts__['test']:
comments.append('State {0} will create '
'the disk group #{1}; cache disk: \'{2}\', '
'capacity disk(s): {3}.'
.format(name, idx, cache_disk_display,
', '.join(['\'{0}\''.format(a) for a in
capacity_disk_displays])))
log.info(comments[-1])
changes = True
continue
try:
__salt__['vsphere.create_diskgroup'](cache_disk_id,
capacity_disk_ids,
safety_checks=False,
service_instance=si)
except VMwareSaltError as err:
comments.append('Error creating disk group #{0}: '
'{1}.'.format(idx, err))
log.error(comments[-1])
errors = True
continue
comments.append('Created disk group #\'{0}\'.'.format(idx))
log.info(comments[-1])
diskgroup_changes[six.text_type(idx)] = \
{'new': {'cache': cache_disk_display,
'capacity': capacity_disk_displays}}
changes = True
continue
# The diskgroup exists; checking the capacity disks
log.debug('Disk group #%s exists. Checking capacity disks: %s.',
idx, capacity_disk_displays)
existing_diskgroup = \
cache_disk_to_existing_diskgroup_map.get(cache_disk_id)
existing_capacity_disk_displays = \
['{0} (id:{1})'.format([d['scsi_address'] for d in host_disks
if d['id'] == disk_id][0], disk_id)
for disk_id in existing_diskgroup['capacity_disks']]
# Populate added disks and removed disks and their displays
added_capacity_disk_ids = []
added_capacity_disk_displays = []
removed_capacity_disk_ids = []
removed_capacity_disk_displays = []
for disk_id in capacity_disk_ids:
if disk_id not in existing_diskgroup['capacity_disks']:
disk_scsi_addr = [d['scsi_address'] for d in host_disks
if d['id'] == disk_id][0]
added_capacity_disk_ids.append(disk_id)
added_capacity_disk_displays.append(
'{0} (id:{1})'.format(disk_scsi_addr, disk_id))
for disk_id in existing_diskgroup['capacity_disks']:
if disk_id not in capacity_disk_ids:
disk_scsi_addr = [d['scsi_address'] for d in host_disks
if d['id'] == disk_id][0]
removed_capacity_disk_ids.append(disk_id)
removed_capacity_disk_displays.append(
'{0} (id:{1})'.format(disk_scsi_addr, disk_id))
log.debug('Disk group #%s: existing capacity disk ids: %s; added '
'capacity disk ids: %s; removed capacity disk ids: %s',
idx, existing_capacity_disk_displays,
added_capacity_disk_displays, removed_capacity_disk_displays)
#TODO revisit this when removing capacity disks is supported
if removed_capacity_disk_ids:
comments.append(
'Error removing capacity disk(s) {0} from disk group #{1}; '
'operation is not supported.'
''.format(', '.join(['\'{0}\''.format(id) for id in
removed_capacity_disk_displays]), idx))
log.error(comments[-1])
errors = True
continue
if added_capacity_disk_ids:
# Capacity disks need to be added to disk group
# Building a string representation of the capacity disks
# that need to be added
s = ', '.join(['\'{0}\''.format(id) for id in
added_capacity_disk_displays])
if __opts__['test']:
comments.append('State {0} will add '
'capacity disk(s) {1} to disk group #{2}.'
''.format(name, s, idx))
log.info(comments[-1])
changes = True
continue
try:
__salt__['vsphere.add_capacity_to_diskgroup'](
cache_disk_id,
added_capacity_disk_ids,
safety_checks=False,
service_instance=si)
except VMwareSaltError as err:
comments.append('Error adding capacity disk(s) {0} to '
'disk group #{1}: {2}.'.format(s, idx, err))
log.error(comments[-1])
errors = True
continue
com = ('Added capacity disk(s) {0} to disk group #{1}'
''.format(s, idx))
log.info(com)
comments.append(com)
diskgroup_changes[six.text_type(idx)] = \
{'new': {'cache': cache_disk_display,
'capacity': capacity_disk_displays},
'old': {'cache': cache_disk_display,
'capacity': existing_capacity_disk_displays}}
changes = True
continue
# No capacity needs to be added
s = ('Disk group #{0} is correctly configured. Nothing to be done.'
''.format(idx))
log.info(s)
comments.append(s)
__salt__['vsphere.disconnect'](si)
#Build the final return message
result = (True if not (changes or errors) else # no changes/errors
None if __opts__['test'] else # running in test mode
False if errors else True) # found errors; defaults to True
ret.update({'result': result,
'comment': '\n'.join(comments),
'changes': diskgroup_changes})
return ret
@depends(HAS_PYVMOMI)
@depends(HAS_JSONSCHEMA)
def host_cache_configured(name, enabled, datastore, swap_size='100%',
dedicated_backing_disk=False,
erase_backing_disk=False):
'''
Configures the host cache used for swapping.
It will do the following:
1. Checks if backing disk exists
2. Creates the VMFS datastore if doesn't exist (datastore partition will be
created and use the entire disk)
3. Raises an error if ``dedicated_backing_disk`` is ``True`` and partitions
already exist on the backing disk
4. Configures host_cache to use a portion of the datastore for caching
(either a specific size or a percentage of the datastore)
Examples
Percentage swap size (can't be 100%)
.. code:: python
{
'enabled': true,
'datastore': {
'backing_disk_scsi_addr': 'vmhba0:C0:T0:L0',
'vmfs_version': 5,
'name': 'hostcache'
}
'dedicated_backing_disk': false
'swap_size': '98%',
}
Fixed sized swap size
.. code:: python
{
'enabled': true,
'datastore': {
'backing_disk_scsi_addr': 'vmhba0:C0:T0:L0',
'vmfs_version': 5,
'name': 'hostcache'
}
'dedicated_backing_disk': true
'swap_size': '10GiB',
}
name
Mandatory state name.
enabled
Specifies whether the host cache is enabled.
datastore
Specifies the host cache datastore.
swap_size
Specifies the size of the host cache swap. Can be a percentage or a
value in GiB. Default value is ``100%``.
dedicated_backing_disk
Specifies whether the backing disk is dedicated to the host cache which
means it must have no other partitions. Default is False
erase_backing_disk
Specifies whether to erase all partitions on the backing disk before
the datastore is created. Default value is False.
'''
log.trace('enabled = %s', enabled)
log.trace('datastore = %s', datastore)
log.trace('swap_size = %s', swap_size)
log.trace('erase_backing_disk = %s', erase_backing_disk)
# Variable used to return the result of the invocation
proxy_details = __salt__['esxi.get_details']()
hostname = proxy_details['host'] if not proxy_details.get('vcenter') \
else proxy_details['esxi_host']
log.trace('hostname = %s', hostname)
log.info('Running host_cache_swap_configured for host \'%s\'', hostname)
ret = {'name': hostname,
'comment': 'Default comments',
'result': None,
'changes': {}}
result = None if __opts__['test'] else True # We assume success
needs_setting = False
comments = []
changes = {}
si = None
try:
log.debug('Validating host_cache_configured input')
schema = HostCacheSchema.serialize()
try:
jsonschema.validate({'enabled': enabled,
'datastore': datastore,
'swap_size': swap_size,
'erase_backing_disk': erase_backing_disk},
schema)
except jsonschema.exceptions.ValidationError as exc:
raise InvalidConfigError(exc)
m = re.match(r'(\d+)(%|GiB)', swap_size)
swap_size_value = int(m.group(1))
swap_type = m.group(2)
log.trace('swap_size_value = %s; swap_type = %s', swap_size_value, swap_type)
si = __salt__['vsphere.get_service_instance_via_proxy']()
host_cache = __salt__['vsphere.get_host_cache'](service_instance=si)
# Check enabled
if host_cache['enabled'] != enabled:
changes.update({'enabled': {'old': host_cache['enabled'],
'new': enabled}})
needs_setting = True
# Check datastores
existing_datastores = None
if host_cache.get('datastore'):
existing_datastores = \
__salt__['vsphere.list_datastores_via_proxy'](
datastore_names=[datastore['name']],
service_instance=si)
# Retrieve backing disks
existing_disks = __salt__['vsphere.list_disks'](
scsi_addresses=[datastore['backing_disk_scsi_addr']],
service_instance=si)
if not existing_disks:
raise VMwareObjectRetrievalError(
'Disk with scsi address \'{0}\' was not found in host \'{1}\''
''.format(datastore['backing_disk_scsi_addr'], hostname))
backing_disk = existing_disks[0]
backing_disk_display = '{0} (id:{1})'.format(
backing_disk['scsi_address'], backing_disk['id'])
log.trace('backing_disk = %s', backing_disk_display)
existing_datastore = None
if not existing_datastores:
# Check if disk needs to be erased
if erase_backing_disk:
if __opts__['test']:
comments.append('State {0} will erase '
'the backing disk \'{1}\' on host \'{2}\'.'
''.format(name, backing_disk_display,
hostname))
log.info(comments[-1])
else:
# Erase disk
__salt__['vsphere.erase_disk_partitions'](
disk_id=backing_disk['id'], service_instance=si)
comments.append('Erased backing disk \'{0}\' on host '
'\'{1}\'.'.format(backing_disk_display,
hostname))
log.info(comments[-1])
# Create the datastore
if __opts__['test']:
comments.append('State {0} will create '
'the datastore \'{1}\', with backing disk '
'\'{2}\', on host \'{3}\'.'
''.format(name, datastore['name'],
backing_disk_display, hostname))
log.info(comments[-1])
else:
if dedicated_backing_disk:
# Check backing disk doesn't already have partitions
partitions = __salt__['vsphere.list_disk_partitions'](
disk_id=backing_disk['id'], service_instance=si)
log.trace('partitions = %s', partitions)
# We will ignore the mbr partitions
non_mbr_partitions = [p for p in partitions
if p['format'] != 'mbr']
if non_mbr_partitions:
raise VMwareApiError(
'Backing disk \'{0}\' has unexpected partitions'
''.format(backing_disk_display))
__salt__['vsphere.create_vmfs_datastore'](
datastore['name'], existing_disks[0]['id'],
datastore['vmfs_version'], service_instance=si)
comments.append('Created vmfs datastore \'{0}\', backed by '
'disk \'{1}\', on host \'{2}\'.'
''.format(datastore['name'],
backing_disk_display, hostname))
log.info(comments[-1])
changes.update(
{'datastore':
{'new': {'name': datastore['name'],
'backing_disk': backing_disk_display}}})
existing_datastore = \
__salt__['vsphere.list_datastores_via_proxy'](
datastore_names=[datastore['name']],
service_instance=si)[0]
needs_setting = True
else:
# Check datastore is backed by the correct disk
if not existing_datastores[0].get('backing_disk_ids'):
raise VMwareSaltError('Datastore \'{0}\' doesn\'t have a '
'backing disk'
''.format(datastore['name']))
if backing_disk['id'] not in \
existing_datastores[0]['backing_disk_ids']:
raise VMwareSaltError(
'Datastore \'{0}\' is not backed by the correct disk: '
'expected \'{1}\'; got {2}'
''.format(
datastore['name'], backing_disk['id'],
', '.join(
['\'{0}\''.format(disk) for disk in
existing_datastores[0]['backing_disk_ids']])))
comments.append('Datastore \'{0}\' already exists on host \'{1}\' '
'and is backed by disk \'{2}\'. Nothing to be '
'done.'.format(datastore['name'], hostname,
backing_disk_display))
existing_datastore = existing_datastores[0]
log.trace('existing_datastore = %s', existing_datastore)
log.info(comments[-1])
if existing_datastore:
# The following comparisons can be done if the existing_datastore
# is set; it may not be set if running in test mode
#
# We support percent, as well as MiB, we will convert the size
# to MiB, multiples of 1024 (VMware SDK limitation)
if swap_type == '%':
# Percentage swap size
# Convert from bytes to MiB
raw_size_MiB = (swap_size_value/100.0) * \
(existing_datastore['capacity']/1024/1024)
else:
raw_size_MiB = swap_size_value * 1024
log.trace('raw_size = %sMiB', raw_size_MiB)
swap_size_MiB = int(raw_size_MiB/1024)*1024
log.trace('adjusted swap_size = %sMiB', swap_size_MiB)
existing_swap_size_MiB = 0
m = re.match(r'(\d+)MiB', host_cache.get('swap_size')) if \
host_cache.get('swap_size') else None
if m:
# if swap_size from the host is set and has an expected value
# we are going to parse it to get the number of MiBs
existing_swap_size_MiB = int(m.group(1))
if not existing_swap_size_MiB == swap_size_MiB:
needs_setting = True
changes.update(
{'swap_size':
{'old': '{}GiB'.format(existing_swap_size_MiB/1024),
'new': '{}GiB'.format(swap_size_MiB/1024)}})
if needs_setting:
if __opts__['test']:
comments.append('State {0} will configure '
'the host cache on host \'{1}\' to: {2}.'
''.format(name, hostname,
{'enabled': enabled,
'datastore_name': datastore['name'],
'swap_size': swap_size}))
else:
if (existing_datastore['capacity'] / 1024.0**2) < \
swap_size_MiB:
raise ArgumentValueError(
'Capacity of host cache datastore \'{0}\' ({1} MiB) is '
'smaller than the required swap size ({2} MiB)'
''.format(existing_datastore['name'],
existing_datastore['capacity'] / 1024.0**2,
swap_size_MiB))
__salt__['vsphere.configure_host_cache'](
enabled,
datastore['name'],
swap_size_MiB=swap_size_MiB,
service_instance=si)
comments.append('Host cache configured on host '
'\'{0}\'.'.format(hostname))
else:
comments.append('Host cache on host \'{0}\' is already correctly '
'configured. Nothing to be done.'.format(hostname))
result = True
__salt__['vsphere.disconnect'](si)
log.info(comments[-1])
ret.update({'comment': '\n'.join(comments),
'result': result,
'changes': changes})
return ret
except CommandExecutionError as err:
log.error('Error: %s.', err)
if si:
__salt__['vsphere.disconnect'](si)
ret.update({
'result': False if not __opts__['test'] else None,
'comment': '{}.'.format(err)})
return ret
def _lookup_syslog_config(config):
'''
Helper function that looks up syslog_config keys available from
``vsphere.get_syslog_config``.
'''
lookup = {'default-timeout': 'Default Network Retry Timeout',
'logdir': 'Local Log Output',
'default-size': 'Local Logging Default Rotation Size',
'logdir-unique': 'Log To Unique Subdirectory',
'default-rotate': 'Local Logging Default Rotations',
'loghost': 'Remote Host'}
return lookup.get(config)
def _strip_key(key_string):
'''
Strips an SSH key string of white space and line endings and returns the new string.
key_string
The string to be stripped.
'''
key_string.strip()
key_string.replace('\n', '')
key_string.replace('\r\n', '')
return key_string