corosync
: Configures the Pacemaker+Corosync stack to provide high-availability.corosync::params
: Configures sane defaults based on the operating system.corosync::qdevice
: Performs basic initial configuration of the qdevice daemon on a node.corosync::reprobe
: Triggers re-probe for changes any of the native cs_* types.
corosync::service
: Declare services within /etc/corosync/service.d/ (Corosync 1.x)
cs_clone
: Type for manipulating corosync/pacemaker resource clone. More information on Corosync/Pacemaker colocation can be found here: * http://www.ccs_colocation
: Type for manipulating corosync/pacemaker colocation. Colocation is the grouping together of a set of primitives so that they travel togethercs_commit
: Final commit statement which triggers the synchronous application of all primitive changes which reference this CIB. Do not generate more thacs_group
: Type for manipulating Corosync/Pacemaker group entries. Groups are a set or resources (primitives) that need to be grouped together. More incs_location
: Type for manipulating corosync/pacemaker resource location. More information on Corosync/Pacemaker colocation can be found here: * http://wwcs_order
: Type for manipulating Corosync/Pacemaker ordering entries. Order entries are another type of constraint that can be put on sets of primitivecs_primitive
: Type for manipulating Corosync/Pacemaker primitives. Primitives are probably the most important building block when creating highly availablcs_property
: Type for manipulating corosync/pacemaker configuration properties. Besides the configuration file that is managed by the module the containscs_rsc_defaults
: Type for manipulating corosync/pacemaker global defaults for resource options. The type is pretty simple interface for setting key/value paircs_shadow
: cs_shadow resources represent a Corosync shadow CIB. Any corosync resources defined with 'cib' set to the title of a cs_shadow resource will
Corosync::ArrayRing
: Custom type for infinitely nestable arraysCorosync::CryptoCipher
: Defines the allowed cipher types for secure corosync communicationCorosync::CryptoHash
: Custom type for possible crypto hashesCorosync::IpStringIp
: Custom type for string <-> array of string variantsCorosync::QuorumAlgorithm
: Custom type for quorumalgorithm enumCorosync::Syslogpriority
: Custom type for syslog priority enum
This class will set up corosync for use by the Puppet Enterprise console to facilitate an active/standby configuration for high availability. It is assumed that this module has been initially ran on a Puppet master with the capabilities of signing certificates to do the initial key generation.
=== Authors
Cody Herriges [email protected]
=== Copyright
Copyright 2012, Puppet Labs, LLC.
class { 'corosync':
enable_secauth => false,
bind_address => '192.168.2.10',
multicast_address => '239.1.1.2',
}
The following parameters are available in the corosync
class:
enable_secauth
authkey_source
authkey
crypto_hash
crypto_cipher
config_validate_cmd
threads
bind_address
pcs_version
port
multicast_address
unicast_addresses
force_online
check_standby
log_timestamp
log_file
log_file_name
debug
log_stderr
syslog_priority
log_function_name
rrp_mode
netmtu
ttl
vsftype
package_corosync
package_pacemaker
package_fence_agents
packageopts_corosync
packageopts_crmsh
packageopts_pacemaker
packageopts_pcs
packageopts_fence_agents
ensure_corosync
ensure_crmsh
ensure_pacemaker
ensure_pcs
ensure_fence_agents
set_votequorum
votequorum_expected_votes
quorum_members
quorum_members_ids
quorum_members_names
token
token_retransmits_before_loss_const
compatibility
enable_corosync_service
manage_corosync_service
enable_pacemaker_service
manage_pacemaker_service
enable_pcsd_service
manage_pcsd_service
manage_pcsd_auth
manage_pcsd_auth_node
sensitive_hacluster_password
sensitive_hacluster_hash
manage_quorum_device
quorum_device_host
quorum_device_algorithm
package_quorum_device
sensitive_quorum_device_password
cluster_name
join
consensus
ip_version
clear_node_high_bit
max_messages
test_corosync_config
watchdog_device
provider
Data type: Boolean
Controls corosync's ability to authenticate and encrypt multicast messages.
Default value: $corosync::params::enable_secauth
Data type: Enum['file', 'string']
Allows to use either a file or a string as a authkey.
Default value: $corosync::params::authkey_source
Data type: Variant[Stdlib::Filesource,Stdlib::Base64]
Specifies the path to the CA which is used to sign Corosync's certificate if authkey_source is 'file' or a base64 encoded version of the actual authkey if 'string' is used instead.
Default value: $corosync::params::authkey
Data type: Corosync::CryptoHash
Hashing algorithm used by corosync for intra-cluster communication. Valid values are none, md5, sha1, sha256, sha384, and sha512
Default value: 'sha1'
Data type: Corosync::CryptoCipher
Encryption cipher used by corosync for intra-cluster communication. Valid values are none, aes256, aes192, aes128, and 3des
Default value: 'aes256'
Data type: String[1]
Default value: '/usr/bin/env COROSYNC_MAIN_CONFIG_FILE=% /usr/sbin/corosync -t'
Data type: Optional[Integer]
How many threads you are going to let corosync use to encode and decode multicast messages. If you turn off secauth then corosync will ignore threads.
Default value: undef
Data type: Corosync::IpStringIp
The ip address we are going to bind the corosync daemon too. Can be specified as an array to have multiple rings.
Default value: $corosync::params::bind_address
Data type: String
Default value: ''
Data type: Optional[Variant[Stdlib::Port, Array[Stdlib::Port]]]
The UDP port that corosync will use to do its multicast communication. Be aware that corosync used this defined port plus minus one. Can be specified as an array to have multiple rings.
Default value: $corosync::params::port
Data type: Optional[Corosync::IpStringIp]
An IP address that has been reserved for multicast traffic. This is the default way that Corosync accomplishes communication across the cluster. Use 'broadcast' to have broadcast instead Can be specified as an array to have multiple rings (multicast only).
Default value: undef
Data type: Optional[Array]
An array of IP addresses that make up the cluster's members. These are used if you are not able to use multicast on your network and instead opt for the udpu transport. You need a relatively recent version of Corosync to make this possible. You can also have an array of arrays to have multiple rings. In that case, each subarray matches a host IP addresses. As of Corosync 3 knet is the new default which also does not use multicast.
Default value: undef
Data type: Boolean
Boolean parameter specifying whether to force nodes that have been put in standby back online.
Default value: $corosync::params::force_online
Data type: Boolean
Boolean parameter specifying whether puppet should return an error log message if a node is in standby. Useful for monitoring node state.
Default value: $corosync::params::check_standby
Data type: Boolean
Boolean parameter specifying whether a timestamp should be placed on all log messages.
Default value: $corosync::params::log_timestamp
Data type: Boolean
Boolean parameter specifying whether Corosync should produce debug output in a logfile.
Default value: $corosync::params::log_file
Data type: Optional[Stdlib::Absolutepath]
Absolute path to the logfile Corosync should use when $log_file
(see
above) is true.
Default value: undef
Data type: Boolean
Boolean parameter specifying whether Corosync should produce debug output in its logs.
Default value: $corosync::params::debug
Data type: Boolean
Boolean parameter specifying whether Corosync should log errors to stderr.
Default value: $corosync::params::log_stderr
Data type: Corosync::SyslogPriority
String parameter specifying the minimal log level for Corosync syslog messages. Allowed values: debug|info|notice|warning|err|emerg.
Default value: $corosync::params::syslog_priority
Data type: Boolean
Boolean parameter specifying whether Corosync should log called function names to.
Default value: $corosync::params::log_function_name
Data type: Optional[Enum['none', 'active', 'passive']]
Mode of redundant ring. May be none, active, or passive.
Default value: undef
Data type: Optional[Integer]
This specifies the network maximum transmit unit.
Default value: undef
Data type: Optional[Integer[0,255]]
Time To Live.
Default value: undef
Data type: Optional[Enum['ykd', 'none']]
Virtual synchrony filter type.
Default value: undef
Data type: Boolean
Define if package corosync should be managed.
Default value: $corosync::params::package_corosync
Data type: Boolean
Define if package pacemaker should be managed.
Default value: $corosync::params::package_pacemaker
Data type: Boolean
Define if package fence-agents should be managed. Default (Red Hat based): true Default (otherwise): false
Default value: false
Data type: Optional[Array[String[1]]]
Additional install-options for the corosync package resource. Default: undef
Default value: $corosync::params::package_install_options
Data type: Optional[Array[String[1]]]
Additional install-options for the crmsh package resource. Default: undef
Default value: $corosync::params::package_install_options
Data type: Optional[Array[String[1]]]
Additional install-options for the pacemaker package resource. Default: undef
Default value: $corosync::params::package_install_options
Data type: Optional[Array[String[1]]]
Additional install-options for the pcs package resource. Default: undef
Default value: $corosync::params::package_install_options
Data type: Optional[Array[String[1]]]
Additional install-options for the pcs package resource. Default: undef
Default value: $corosync::params::package_install_options
Data type: String[1]
Define what version of the corosync package should be installed. Default: 'present'
Default value: $corosync::params::ensure_corosync
Data type: String[1]
Define what version of the crmsh package should be installed. Default: 'present'
Default value: $corosync::params::ensure_crmsh
Data type: String[1]
Define what version of the pacemaker package should be installed. Default: 'present'
Default value: $corosync::params::ensure_pacemaker
Data type: String[1]
Define what version of the pcs package should be installed. Default: 'present'
Default value: $corosync::params::ensure_pcs
Data type: String[1]
Define what version of the fence-agents-all package should be installed. Default: 'present'
Default value: $corosync::params::ensure_fence_agents
Data type: Boolean
Set to true if corosync_votequorum should be used as quorum provider. Default (Red Hat based): true Default (Ubuntu >= 14.04): true Default (otherwise): false
Default value: $corosync::params::set_votequorum
Data type: Optional[Integer]
Overrides the automatic calculation of expected votes which is normally derived from the number of nodes.
Default value: undef
Data type: Array
Array of quorum member hostname. This is required if set_votequorum is set to true. You can also have an array of arrays to have multiple rings. In that case, each subarray matches a member IP addresses.
Default value: ['localhost']
Data type: Optional[Array]
Array of quorum member IDs. Persistent IDs are required for the dynamic config of a corosync cluster and when_set_votequorum is set to true. Should be used only with the quorum_members parameter.
Default value: undef
Data type: Optional[Array]
Array of quorum member names. Persistent names are required when you define IP addresses in quorum_members.
Default value: undef
Data type: Optional[Integer]
Time (in ms) to wait for a token
Default value: undef
Data type: Optional[Integer]
How many token retransmits before forming a new configuration.
Default value: undef
Data type: Optional[String]
Older versions of corosync allowed a config-file directive to indicate backward compatibility. This sets that.
Default value: undef
Data type: Boolean
Whether the module should enable the corosync service.
Default value: $corosync::params::enable_corosync_service
Data type: Boolean
Whether the module should try to manage the corosync service. If set to false, the service will need to be specified in the catalog elsewhere.
Default value: $corosync::params::manage_corosync_service
Data type: Boolean
Whether the module should enable the pacemaker service.
Default value: $corosync::params::enable_pacemaker_service
Data type: Boolean
Whether the module should try to manage the pacemaker service. Default (Red Hat based >= 7): true Default (Ubuntu >= 14.04): true Default (otherwise): false
Default value: $corosync::params::manage_pacemaker_service
Data type: Boolean
Whether the module should enable the pcsd service.
Default value: $corosync::params::enable_pcsd_service
Data type: Boolean
Whether the module should try to manage the pcsd service in addition to the corosync service. pcsd service is the GUI and the remote configuration interface.
Default value: false
Data type: Boolean
This only has an effect when $manage_pcsd_service is enabled. If set, an attempt will be made to authorize pcs on the cluster node determined by manage_pcsd_auth_node. Note that this determination can only be made when the entries in quorum_members match the trusted certnames of the nodes in the environment or the IP addresses of the primary adapters. $sensitive_hacluster_password is mandatory if this parameter is set.
Default value: false
Data type: Enum['first','last']
When managing authorization for PCS this determines which node does the work. Note that only one node 'should' do the work and nodes are chosen by matching local facts to the contents of quorum_members. When manage_pcsd_auth is disabled this parameter has no effect.
Default value: 'first'
Data type: Optional[Sensitive[String]]
When PCS is configured on a RHEL system this directive is used to set the password for the hacluster user. If both $manage_pcsd_service and $manage_pcsd_auth are both set to true the cluster will use this credential to authorize all nodes.
Default value: undef
Data type: Optional[Sensitive[String]]
This parameter expects a valid password hash of sensitive_hacluster_password. If provided, the hash provided the hash will be used to set the password for the hacluster user on each node.
Default value: undef
Data type: Boolean
Enable or disable the addition of a quorum device external to the cluster. This device is used avoid cluster splits typically in conjunction with fencing by providing an external network vote. Additionally, this allows symmentric clusters to continue operation in the event that 50% of their nodes have failed.
Default value: false
Data type: Optional[Stdlib::Fqdn]
The fully qualified hostname of the quorum device. This parameter is mandatory when manage_quorum_device is true.
Default value: undef
Data type: Corosync::QuorumAlgorithm
There are currently two algorithms the quorum device can utilize to determine how its vote should be allocated; Fifty-fifty split and last-man-standing. See the corosync-qdevice man page for details.
Default value: 'ffsplit'
Data type: Optional[String]
The name of the package providing the quorum device functionality. This parameter is mandatory if manage_quorum_device is true.
Default value: $corosync::params::package_quorum_device
Data type: Optional[Sensitive[String]]
The plain text password for the hacluster user on the quorum_device_host. This parameter is mandatory if manage_quorum_device is true.
Default value: undef
Data type: Optional[String[1]]
This specifies the name of cluster and it's used for automatic generating of multicast address.
Default value: undef
Data type: Optional[Integer]
This timeout specifies in milliseconds how long to wait for join messages in the membership protocol.
Default value: undef
Data type: Optional[Integer]
This timeout specifies in milliseconds how long to wait for consensus to be achieved before starting a new round of membership configuration. The minimum value for consensus must be 1.2 * token. This value will be automatically calculated at 1.2 * token if the user doesn't specify a consensus value.
Default value: undef
Data type: Optional[String[1]]
This specifies version of IP to ask DNS resolver for. The value can be one of ipv4 (look only for an IPv4 address) , ipv6 (check only IPv6 address), ipv4-6 (look for all address families and use first IPv4 address found in the list if there is such address, otherwise use first IPv6 address) and ipv6-4 (look for all address families and use first IPv6 address found in the list if there is such address, otherwise use first IPv4 address).
Default (if unspecified) is ipv6-4 for knet and udpu transports and ipv4 for udp.
Default value: undef
Data type: Optional[Enum['yes', 'no']]
This configuration option is optional and is only relevant when no nodeid is specified. Some openais clients require a signed 32 bit nodeid that is greater than zero however by default openais uses all 32 bits of the IPv4 address space when generating a nodeid. Set this option to yes to force the high bit to be zero and therefor ensure the nodeid is a positive signed 32 bit integer. WARNING: The clusters behavior is undefined if this option is enabled on only a subset of the cluster (for example during a rolling upgrade).
Default value: undef
Data type: Optional[Integer]
This constant specifies the maximum number of messages that may be sent by one processor on receipt of the token. The max_messages parameter is limited to 256000 / netmtu to prevent overflow of the kernel transmit buffers.
Default value: undef
Data type: Boolean
Whether we should test new configuration files with corosync -t
.
(requires corosync 2.3.4)
Default value: $corosync::params::test_corosync_config
Data type: Optional[Variant[Stdlib::Absolutepath, Enum['off']]]
Watchdog device to use, for example '/dev/watchdog' or 'off'. Its presence (or lack thereof) shifted with corosync versions.
Default value: undef
Data type: Enum['pcs', 'crm']
What command line utility provides corosync configuration capabilities.
Default value: 'pcs'
Configures sane defaults based on the operating system.
This class performs the configuration of the qdevice daemon on a target node. Note that this requires corosync 2.x and must never be deployed on a node which is actually part of a cluster. Additionally, you will need to open the correct firewall ports for both pcs, and the actual quorum device as shown in the included example.
include firewalld
class { 'corosync::qdevice':
sensitive_hacluster_hash => $sensitive_hacluster_hash,
}
contain 'corosync::qdevice'
# Open the corosync-qnetd port
firewalld::custom_service { 'corosync-qdevice-net':
description => 'Corosync Quorum Net Device Port',
port => [
{
port => '5403',
protocol => 'tcp',
},
],
}
firewalld_service { 'corosync-qdevice-net':
ensure => 'present',
service => 'corosync-qdevice-net',
zone => 'public',
}
# Configure general PCS firewall rules
firewalld_service { 'high-availability':
ensure => 'present',
service => 'high-availability',
zone => 'public',
}
The following parameters are available in the corosync::qdevice
class:
Data type: Optional[Sensitive[String]]
The password hash for the hacluster user on this quorum device node. If omitted, you must create the hacluster user and haclient group yourself. This user is required because pcsd must be used to perform the quorum node configuration.
Default value: undef
Data type: String[1]
Name of the PCS package on this system.
Default value: 'pcs'
Data type: String[1]
Name of the corosync qnetd package for this system.
Default value: 'corosync-qnetd'
Data type: String
What command line utility provides corosync configuration capabilities.
Include this class to reprobe the corosync cluster when there are changes in any of the native cs_* types. Useful for multi-node provisioning when the nodes are not always in a stable state after provisioning.
Copyright 2012 Puppet Labs, LLC.
include corosync::reprobe
Models a Corosync service. Corosync services are plugins that provide functionality for monitoring cluster resources. One of the most common of these plugins being Pacemaker. This is for corosync 1.x!
=== Authors
Cody Herriges [email protected]
=== Copyright
Copyright 2012 Puppet Labs, LLC.
corosync::service { 'pacemaker':
version => '0',
}
The following parameters are available in the corosync::service
defined type:
Data type: String
The namevar in this type is the title you give it when you define a resource instance. It is used for a handful of purposes; defining the name of the config file and the name defined inside the file itself.
Data type: String[1]
Version of the protocol used by this service. This is currently unused.
Type for manipulating corosync/pacemaker resource clone. More information on Corosync/Pacemaker colocation can be found here:
The following properties are available in the cs_clone
type.
Valid values: %r{\d+}
, absent
How many copies of the resource to start. Defaults to the number of nodes in the cluster.
Default value: absent
Valid values: %r{\d+}
, absent
How many copies of the resource can be started on a single node. Defaults to 1.
Default value: absent
Valid values: present
, absent
The basic property that the resource should be in.
Default value: present
Valid values: true
, false
, absent
Does each copy of the clone perform a different function? Allowed values: true, false
Default value: absent
The corosync resource group to be cloned.
Valid values: true
, false
, absent
Changes the behavior of ordering constraints (between clones/masters) so that instances can start/stop as soon as their peer instance has (rather than waiting for every instance of the other clone has). Allowed values: true, false
Default value: absent
Valid values: true
, false
, absent
When stopping or starting a copy of the clone, tell all the other copies beforehand and when the action was successful. Allowed values: true, false
Default value: absent
Valid values: true
, false
, absent
Should the copies be started in series (instead of in parallel). Allowed values: true, false
Default value: absent
The corosync resource primitive to be cloned.
Valid values: true
, false
, absent
If true, clone instances can perform a special role that Pacemaker will manage via the resource agent’s promote and demote actions. The resource agent must support these actions. Allowed values: false, true
Default value: absent
Valid values: %r{\d+}
, absent
If promotable is true, the number of instances that can be promoted at one time across the entire cluster
Default value: absent
Valid values: %r{\d+}
, absent
If promotable is true and globally-unique is false, the number of clone instances can be promoted at one time on a single node
Default value: absent
The following parameters are available in the cs_clone
type.
Corosync applies its configuration immediately. Using a CIB allows you to group multiple primitives and relationships to be applied at once. This can be necessary to insert complex configurations into Corosync correctly.
This parameter sets the CIB this colocation should be created in. A cs_shadow resource with a title of the same name as this value should also be added to your manifest.
namevar
Identifier of the clone entry. This value needs to be unique across the entire Corosync/Pacemaker configuration since it doesn't have the concept of name spaces per type.
The specific backend to use for this cs_clone
resource. You will seldom need to specify this --- Puppet will usually
discover the appropriate provider for your platform.
Type for manipulating corosync/pacemaker colocation. Colocation is the grouping together of a set of primitives so that they travel together when one of them fails. For instance, if a web server vhost is colocated with a specific ip address and the web server software crashes, the ip address with migrate to the new host with the vhost.
More information on Corosync/Pacemaker colocation can be found here:
The following properties are available in the cs_colocation
type.
Valid values: present
, absent
The basic property that the resource should be in.
Default value: present
At least two Pacemaker primitives to be located together. Order of primitives in colocation groups is important. In Pacemaker, a colocation of 2 primitives behaves different than a colocation between more than 2 primitives. Here the behavior is altered to be more consistent. Examples on how to define colocations here:
- 2 primitives: [A, B] will cause A to be located first, and B will be located with A. This is different than how crm configure colocation works, because there [A, B] would mean colocate A with B, thus B should be located first.
- multiple primitives: [A, B, C] will cause A to be located first, B next, and finally C. This is identical to how crm configure colocation works with multiple resources, it will add a colocated set. Property will raise an error if you do not provide an array containing at least two values. Values can be either the name of the primitive, or primitive:role. Notice, we can only interpret colocations of single sets, not multiple sets combined. In Pacemaker speak, this means we can support 'A B C' but not e.g. 'A B (C D) E'. Feel free to contribute a patch for this.
The priority of this colocation. Primitives can be a part of multiple colocation groups and so there is a way to control which primitives get priority when forcing the move of other primitives. This value can be an integer but is often defined as the string INFINITY.
Default value: INFINITY
The following parameters are available in the cs_colocation
type.
Corosync applies its configuration immediately. Using a CIB allows you to group multiple primitives and relationships to be applied at once. This can be necessary to insert complex configurations into Corosync correctly.
This paramater sets the CIB this colocation should be created in. A cs_shadow resource with a title of the same name as this value should also be added to your manifest.
namevar
Identifier of the colocation entry. This value needs to be unique across the entire Corosync/Pacemaker configuration since it doesn't have the concept of name spaces per type.
The specific backend to use for this cs_colocation
resource. You will seldom need to specify this --- Puppet will
usually discover the appropriate provider for your platform.
Final commit statement which triggers the synchronous application of all primitive changes which reference this CIB. Do not generate more than one cs_commit referencing the same CIB for a given cluster!
The following parameters are available in the cs_commit
type.
Name of the CIB to commit. This value defaults to the name of the cs_commit resource.
namevar
Name of the CIB to commit. See the cib parameter for more detail.
The specific backend to use for this cs_commit
resource. You will seldom need to specify this --- Puppet will usually
discover the appropriate provider for your platform.
Type for manipulating Corosync/Pacemaker group entries. Groups are a set or resources (primitives) that need to be grouped together.
More information can be found at the following link:
The following properties are available in the cs_group
type.
Valid values: present
, absent
The basic property that the resource should be in.
Default value: present
An array of primitives to have in this group. Must be listed in the order that you wish them to start.
The following parameters are available in the cs_group
type.
Corosync applies its configuration immediately. Using a CIB allows you to group multiple primitives and relationships to be applied at once. This can be necessary to insert complex configurations into Corosync correctly.
This parameter sets the CIB this order should be created in. A cs_shadow resource with a title of the same name as this value should also be added to your manifest.
namevar
Name identifier of this group entry. This value needs to be unique across the entire Corosync/Pacemaker configuration since it doesn't have the concept of name spaces per type.
The specific backend to use for this cs_group
resource. You will seldom need to specify this --- Puppet will usually
discover the appropriate provider for your platform.
Type for manipulating corosync/pacemaker resource location. More information on Corosync/Pacemaker colocation can be found here:
The following properties are available in the cs_location
type.
Valid values: present
, absent
The basic property that the resource should be in.
Default value: present
The corosync node_name where the resource should be located.
The corosync resource primitive to have a location applied.
Whether Pacemaker should perform resource discovery on this node for the specified resource. It matches the resource-discovery location property in pacemaker
The rules of this location. This is an array of hashes where each hash contains an array of one or more expressions.
Example:
cs_location { 'vip-ping-connected': primitive => 'vip', rules => [ 'vip-ping-exclude-rule' => { 'score' => '-INFINITY', 'expression' => [ { 'attribute' => 'pingd', 'operation' => 'lt', 'value' => '100', }, ], }, 'vip-ping-prefer-rule' => { 'score-attribute' => 'pingd', 'expression' => [ { 'attribute' => 'pingd', 'operation' => 'defined', }, ], }, ], }
The priority of this location. Primitives can be a part of multiple location groups and so there is a way to control which primitives get priority when forcing the move of other primitives. This value can be an integer but is often defined as the string INFINITY.
Default value: INFINITY
The following parameters are available in the cs_location
type.
Corosync applies its configuration immediately. Using a CIB allows you to group multiple primitives and relationships to be applied at once. This can be necessary to insert complex configurations into Corosync correctly.
This paramater sets the CIB this colocation should be created in. A cs_shadow resource with a title of the same name as this value should also be added to your manifest.
namevar
Identifier of the location entry. This value needs to be unique across the entire Corosync/Pacemaker configuration since it doesn't have the concept of name spaces per type.
The specific backend to use for this cs_location
resource. You will seldom need to specify this --- Puppet will
usually discover the appropriate provider for your platform.
Type for manipulating Corosync/Pacemaker ordering entries. Order entries are another type of constraint that can be put on sets of primitives but unlike colocation, order does matter. These designate the order at which you need specific primitives to come into a desired state before starting up a related primitive.
More information can be found at the following link:
The following properties are available in the cs_order
type.
Valid values: present
, absent
The basic property that the resource should be in.
Default value: present
First Corosync primitive. Just like colocation, our primitives for ordering come in pairs but this time order matters so we need to define which primitive starts the desired state change chain.
How to enforce the constraint.
Allowed values:
- Optional: Just a suggestion. Only applies if both resources are executing the specified actions. Any change in state by the first resource will have no effect on the then resource.
- Mandatory: Always. If first does not perform first-action, then will not be allowed to performed then-action. If first is restarted, then (if running) will be stopped beforehand and started afterward.
- Serialize: Ensure that no two stop/start actions occur concurrently for the resources. First and then can start in either order, but one must complete starting before the other can be started. A typical use case is when resource start-up puts a high load on the host.
Default value: Mandatory
The priority of the this ordered grouping. Primitives can be a part
of multiple order groups and so there is a way to control which
primitives get priority when forcing the order of state changes on
other primitives. This value can be an integer but is often defined
as the string INFINITY.
When using pcs as provider this value is not used.
It is generally preferred to use the kind
parameter.
Second Corosync primitive. Our second primitive will move to the desired state after the first primitive.
Boolean specifying if the resources should stop in reverse order. Default value: true.
Default value: true
The following parameters are available in the cs_order
type.
Corosync applies its configuration immediately. Using a CIB allows you to group multiple primitives and relationships to be applied at once. This can be necessary to insert complex configurations into Corosync correctly.
This parameter sets the CIB this order should be created in. A cs_shadow resource with a title of the same name as this value should also be added to your manifest.
namevar
Name identifier of this ordering entry. This value needs to be unique across the entire Corosync/Pacemaker configuration since it doesn't have the concept of name spaces per type.
The specific backend to use for this cs_order
resource. You will seldom need to specify this --- Puppet will usually
discover the appropriate provider for your platform.
Type for manipulating Corosync/Pacemaker primitives. Primitives are probably the most important building block when creating highly available clusters using Corosync and Pacemaker. Each primitive defines an application, ip address, or similar to monitor and maintain. These managed primitives are maintained using what is called a resource agent. These resource agents have a concept of class, type, and subsystem that provides the functionality. Regrettably these pieces of vocabulary clash with those used in Puppet so to overcome the name clashing the property and parameter names have been qualified a bit for clarity.
More information on primitive definitions can be found at the following link:
The following properties are available in the cs_primitive
type.
Valid values: present
, absent
The basic property that the resource should be in.
Default value: present
A hash of metadata for the primitive. A primitive can have a set of metadata that doesn't affect the underlying Corosync type/provider but affect that concept of a resource. This metadata is similar to Puppet's resources resource and some meta-parameters, they change resource behavior but have no affect of the data that is synced or manipulated.
Default value: Hash.new
A hash of operations for the primitive. Operations defined in a primitive are little more predictable as they are commonly things like monitor or start and their values are in seconds. Since each resource agent can define its own set of operations we are going to defer again and just accept a hash. There maybe room to model this one but it would require a review of all resource agents to see if each operation is valid.
Default value: Hash.new
A hash of params for the primitive. Parameters in a primitive are used by the underlying resource agent, each class using them slightly differently. In ocf scripts they are exported and pulled into the script as variables to be used. Since the list of these parameters are completely arbitrary and validity not enforced we simply defer defining a model and just accept a hash.
Default value: Hash.new
A hash of utilization attributes for the primitive. If nodes are also configured with available resources, and Pacemaker's placement strategy is set appropriately, then Pacemaker can place primitives on nodes only where resources are available.
See the Pacemaker documentation:
http://clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/ch11.html
Default value: Hash.new
The following parameters are available in the cs_primitive
type.
Corosync applies its configuration immediately. Using a CIB allows you to group multiple primitives and relationships to be applied at once. This can be necessary to insert complex configurations into Corosync correctly.
This parameter sets the CIB this primitive should be created in. A cs_shadow resource with a title of the same name as this value should also be added to your manifest.
namevar
Name identifier of primitive. This value needs to be unique across the entire Corosync/Pacemaker configuration since it doesn't have the concept of name spaces per type.
Corosync class of the primitive. Examples of classes are lsb or ocf. Lsb functions a lot like the init provider in Puppet for services, an init script is ran periodically on each host to identify status, or to start and stop a particular application. Ocf of the other hand is a script with meta-data and structure that is specific to Corosync and Pacemaker.
Corosync primitive type. Type generally matches to the specific 'thing' your managing, i.e. ip address or vhost. Though, they can be completely arbitrarily named and manage any number of underlying applications or resources.
Corosync primitive provider. All resource agents used in a primitive have something that provides them to the system, be it the Pacemaker or redhat plugins...they're not always obvious though so currently you're left to understand Corosync enough to figure it out. Usually, if it isn't obvious it is because there is only one provider for the resource agent.
To find the list of providers for a resource agent run the following from the command line has Corosync installed:
crm configure ra providers <ra> <class>
The specific backend to use for this cs_primitive
resource. You will seldom need to specify this --- Puppet will
usually discover the appropriate provider for your platform.
Metadata options that should not be managed by Puppet. Examples: ['target-role', 'is-managed']
Default value: Array.new
Type for manipulating corosync/pacemaker configuration properties. Besides the configuration file that is managed by the module the contains all these related Corosync types and providers, there is a set of cluster properties that can be set and saved inside the CIB (A CIB being a set of configuration that is synced across the cluster, it can be exported as XML for processing and backup). The type is pretty simple interface for setting key/value pairs or removing them completely. Removing them will result in them taking on their default value.
More information on cluster properties can be found here:
P.S Looked at generating a type dynamically from the cluster's property meta-data that would result in a single type with puppet type properties of every cluster property...may still do so in a later iteration.
The following properties are available in the cs_property
type.
Valid values: present
, absent
The basic property that the resource should be in.
Default value: present
Value of the property. It is expected that this will be a single value but we aren't validating string vs. integer vs. boolean because cluster properties can range the gambit.
The following parameters are available in the cs_property
type.
Corosync applies its configuration immediately. Using a CIB allows you to group multiple primitives and relationships to be applied at once. This can be necessary to insert complex configurations into Corosync correctly.
This parameter sets the CIB this parameter should be created in. A cs_shadow resource with a title of the same name as this value should also be added to your manifest.
namevar
Name identifier of this property. Simply the name of the cluster property. Happily most of these are unique.
The specific backend to use for this cs_property
resource. You will seldom need to specify this --- Puppet will
usually discover the appropriate provider for your platform.
Valid values: true
, false
, yes
, no
Whether to replace a property that already exists on the cluster
whose value doesn't match what the value
attribute specifies. Setting
this to false allows cs_property resources to initialize properties without
overwriting future changes. Defaults to true
.
Default value: true
Type for manipulating corosync/pacemaker global defaults for resource options. The type is pretty simple interface for setting key/value pairs or removing them completely. Removing them will result in them taking on their default value.
More information on resource defaults can be found here:
- http://clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/s-resource-defaults.html
- http://clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/s-resource-options.html
The following properties are available in the cs_rsc_defaults
type.
Valid values: present
, absent
The basic property that the resource should be in.
Default value: present
Value of the property. It is expected that this will be a single value but we aren't validating string vs. integer vs. boolean because resource options can range the gambit.
The following parameters are available in the cs_rsc_defaults
type.
Corosync applies its configuration immediately. Using a CIB allows you to group multiple primitives and relationships to be applied at once. This can be necessary to insert complex configurations into Corosync correctly.
This parameter sets the CIB this rsc_defaults should be created in. A cs_shadow resource with a title of the same name as this value should also be added to your manifest.
namevar
Name identifier of this property. Simply the name of the resource option. Happily most of these are unique.
The specific backend to use for this cs_rsc_defaults
resource. You will seldom need to specify this --- Puppet will
usually discover the appropriate provider for your platform.
cs_shadow resources represent a Corosync shadow CIB. Any corosync resources defined with 'cib' set to the title of a cs_shadow resource will not become active until all other resources with the same cib value have also been applied.
The following properties are available in the cs_shadow
type.
Implementation detail. DO NOT SET DIRECTLY.
Default value: latest
The following parameters are available in the cs_shadow
type.
Valid values: true
, false
, yes
, no
Whether to generate a cs_commit or not. Can be used to create shadow CIB without committing them.
Default value: true
namevar
Name of the CIB to begin tracking changes against.
The specific backend to use for this cs_shadow
resource. You will seldom need to specify this --- Puppet will usually
discover the appropriate provider for your platform.
Custom type for infinitely nestable arrays
Alias of
Variant[Array[Stdlib::IP::Address], Array[
Array[Stdlib::IP::Address]
]]
Defines the allowed cipher types for secure corosync communication
Alias of Enum['aes256', 'aes192', 'aes128', '3des']
Custom type for possible crypto hashes
Alias of Enum['md5', 'sha1', 'sha256', 'sha384', 'sha512']
Custom type for string <-> array of string variants
Alias of
Variant[Stdlib::IP::Address, Array[
Stdlib::IP::Address
]]
Custom type for quorumalgorithm enum
Alias of Enum['ffsplit', 'lms']
Custom type for syslog priority enum
Alias of Enum['debug', 'info', 'notice', 'warning', 'err', 'alert', 'emerg', 'crit']