RHEL5 cluster.conf Schema Description

Tag: <cluster>
Attributes:

• name (required): The name of the cluster you have configured.
• config_version (required): Specifies a revision number for the cluster.conf file.
  In order to successfully propagate a new configuration file,
  this parameter must be higher than current value, i.e. monotonically increasing order.
• Alias: Pretty name for cluster that is not parsed by cluster
  tag software; only the gui uses this value.

Tag: <fence_xvmd> Fence_xvm daemon startup flag
Note: This tag is optional. fence_xvmd is an I/O Fencing host which resides on dom0 and is used in conjunction with the fence_xvm fencing agent. Together, these two programs can be used to fence can be used to fence Xen virtual machines which are part of a cluster. There is a requirement that the parent dom0s are also a part of their own CMAN/OpenAIS based cluster, and that the dom0 cluster does not share any members with the domU cluster. Furthermore, the dom0 cluster is required to have fencing if domU recovery is expected to be automatic.
Parent Tag: <cluster>
Attributes:

• family (Optional): IP family to use (auto, ipv4, or ipv6; default = auto)
• multicast_address (Optional): Multicast address to listen on (default=225.0.0.12 for ipv4, ff02::3:1 for ipv6)
• port (Optional): Port to use (default=1229)
• auth (Optional): Authentication type (none, sha1, sha256, sha512; default=sha256). This controls the authentication mechanism used to authenticate clients. The three SHA hashes use a key which must be shared between both the Xen virtual machines and the host domain-0 cluster. The three SHA authentication mecha- nisms use a simple bidirectional challenge-response based on pseudo- random number generation and a shared private key.
• hash (Optional): Packet hash type (none, sha1, sha256, sha512; default=sha256). This controls the hashing mechanism used to authenticate fencing requests. The three SHA hashes use a key which must be shared between both the Xen virtual machines and the host domain-0 cluster.
• key_file (Optional): Use the specified key file for packet hashing / SHA authentication. When both the hash type and the authentication type are set to "none" (or not used), this parameter is ignored.
• use_uuid=1 (Optional): Fence by UUID instead of Xen Domain name.

Tag: <totem> OpenAIS msg transport protocol
Parent Tag: <cluster>
Attributes:

• token (Optional): This timeout specifies in milliseconds until a token loss is declared after not receiving a token. This is the time spent detecting a failure of a processor in the current configuration. Reforming a new configuration takes about 50 milliseconds in addition to this timeout.

  The default is 10000 milliseconds.

• token_retransmits_before_loss_const (Optional): This timeout specifies in milliseconds after how long before receiving a token the token is retransmitted. This will be automatically calculated if token is modified. It is not recommended to alter this value without guidance from the openais community.

  The default is 20 milliseconds.

• join (Optional): This timeout specifies in milliseconds how long to wait for join messages in the membership protocol.

  The default is 60 milliseconds.

• consensus (Optional): This timeout specifies in milliseconds how long to wait for consensus to be achieved before starting a new round of membership configuration.

  The default is 4800 milliseconds.

Tag: <cman> CMAN service configuration
Parent Tag: <cluster>
Attributes:

Tag: <multicast>
Parent Tag: <cman>
Attributes:

• addr: Address for multicasting.
Cman can be configured to use multicast instead
of broadcast (broadcast is used by default if no
multicast parameters are given.) Example: ==== <multicast addr="224.0.0.1"/> ====

Tag: <fence_daemon> Fence Daemon Configuration
Parent Tag: <cluster>
Attributes:

• post_join_delay: The number of seconds the daemon will wait before
  fencing any victims after a node joins the domain.
• post_fail_delay: The number of seconds the daemon will wait before
  fencing any victims after a domain member fails.
• clean_start: Used to prevent any start up fencing the daemon might
  do. It indicates that the daemon should assume all nodes
  are in a clean state to start.
Example: ==== <fence_daemon post_fail_delay="0" post_join_delay="3"/> <fence_daemon clean_start="0"/> ====

A Note On Fencing

Fencing is specified within the cluster.conf file in two places. The first place is within the <fencedevices> tag. Any device used for fencing a node must be defined here as a <fencedevice> first. This applies to power switches (APC, WTI, etc.) with multiple ports that are able to fence multiple cluster nodes, as well as fabric switches and baseboard management fence strategies (iLO, RSA, IPMI, Drac, etc.) that are usually 1 to 1 in nature; that is, one specified fence device is able to fence only one node. After defining the fence devices to be used in the cluster, it is necessary to associate the fence device listings with specific cluster nodes. The second place that fencing is specified within cluster.conf is within the <clusternode> tag. Beneath the <clusternode> tag, is a <fence> tag. Beneath the <fence> tag is one or more <method> tag sets. Within a <method> tag set, is a <device> tag set. This is where the actual association between <fencedevice> and node takes place. A <device> tag has a required "name" attribute that refers to the name of one of the <fencedevice>'s specified in the <fencedevices> section of cluster.conf. More about <method> blocks: A method block is like a fence level. If a primary fence method is selected, yet the user wants to define a backup method in case the first fence method fails, this is done by defining two <metho>d blocks for a cluster node, each with a unique name parameter. The fence daemon will call each fence method in the order they are specified under the <clusternode><fence> tag set. Fence specification within cluster.conf offers one other feature for customizing fence action. Within a <method> block, it is allowable to list more than one <device>. This is useful when fencing a node with redundant power supplies, for example. The fence daemon will run the agent for each device listed within a <method> block before determining success or failure.

Tag: <fencedevices> Contains all fencing device information.
Parent Tag: <cluster>
Attributes: None

Tag: <fencedevice> Information about one fence device in particular.
Parent Tag: <fencedevices>
Attributes:

• name (required by ALL fence devices): Name of the fence device.
• agent (required by ALL fence devices): Specifies fence agent to be used. See
  agent list in Table 1 below.
• The following tables list attributes depending on which fence agent is used (all of these attributes listed below are attributes):
  

  Type: APC Power Switch

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_apc ipaddr IP address of the device. login login name for device. passwd password.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above port switch outlet port switch Optional: switch number when ganging Masterswitch Plus switches option Action required. 'Reboot' (default action if this attr is not present) 'Off' or 'On'

  Type: Brocade Fabric Switch

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_brocade ipaddr IP address of the device. login login name for device. passwd password.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above port switch port option Action required. 'disable' (default action if this attr is not present) or 'enable'

  Type: McData SAN Switch

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_mcdata ipaddr IP address of the device. login login name for device. passwd password.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above port switch port option Action required. 'disable' (default action if this attr is not present) or 'enable'

  Type: QLogic SANBox2

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_sanbox2 ipaddr IP address of the device. login login name for device. passwd password.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above port switch port option Action required. 'disable' (default action if this attr is not present) or 'enable'

  Type: IBM Blade Center

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_bladecenter ipaddr IP address of the device. login login name for device. passwd password.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above blade the blade to operate on option Action required. 'disable' (default action if this attr is not present) or 'enable'

  Type: Bull PAP

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_bullpap ipaddr IP address of the device. login login name for device. passwd password.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above domain Domain of the BullPAP system to power cycle option Action required. 'reboot' (default action if this attr is not present) 'on' or 'off'

  Type: IPMI Lan

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_ipmilan ipaddr IP address of the device. login login name for device. passwd password. auth Authentication Type: none, 'password', 'md2', or 'md5' lanplus Use lanplus: 'True' or '1'; leave out for false (only in RHEL4.5 CS and later)

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above option Action required. 'reboot' (default action if this attr is not present) 'on' or 'off'

  Type: WTI Power Switch

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_wti ipaddr IP address of the device. passwd password.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above port The switch outlet port to power cycle

  Type: Vixel SAN Switch

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_vixel ipaddr IP address of the device. passwd password.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above port The switch port to remove zoning from

  Type: HP ilo (Integrated Lights Out)

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_ilo hostname Hostname or IP Address of the device. login Login for the device. passwd password for the device. ribcl RIBCL protocol version to use. Default action if this attr is not present is to autodetect

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above action Action required. 'reboot' (default action if this attr is not present) 'on' or 'off' force Optional parameter. If set to '1' (force='1'), status will not be initially checked before fencing. The use of this attribute significantly speeds up fencing on this device type.

  Type: Global Network Block Device (GNBD)

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_gnbd servers A whitespace separated list of servers

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above

  Type: Egenera SAN Controller

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_egenera cserver The cserver to ssh to...the cserver can be of the form user@hostname to log in as a particular user esh The path to the esh command on the cserver (default is /opt/pan-mgr/bin/esh if this attr is not present)

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above lpan The lpan to use pserver The pserver to fence for this node action Action required. 'reboot' (default action if this attr is not present) 'on' or 'off'

  Type: IBM RSA II

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_rsa ipaddr IP Address or Hostname of the device. login Login for the device. passwd password for the device.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above option Action required. 'Reboot' (default action if this attr is not present) 'On' or 'Off'

  Type: RPS10 Serial Switch

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_rps10 device Device name. port Port to fence.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above

  Type: Dell DRAC

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_drac ipaddr Hostname or IP Address of the device. login Login for the device. passwd password for the device. drac_version Force fence agent to use a particular version of DRAC firmware. The default action when not including this attr is to auto-detect cmd_prompt Optional: Force fence_drac to use this value as the command prompt.

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above action Action required. 'reboot' (default action if this attr is not present) 'on' or 'off' modulename Optional: used when employing DRAC/MC multi-chassis version

  Type: Manual Fencing

  <fencedevice> attributes

  name reference name for this device within the conf file agent fence_manual

  <clusternode> <method> <device> attributes

  name reference to the fence device name attribute above

  Example: ==== <fencedevices> <fencedevice agent="fence_wti" ipaddr="gfsrack" name="wti" passwd="pa$$word"/> <fencedevice agent="fence_wti" ipaddr="rack007" name="wti2" passwd="pa$$word"/> </fencedevices> ====

  Tag: <clusternodes> Cluster Nodes Configuration: contains 1 or more <clusternode> tags.>br/> Parent Tag: <cluster>
  Attributes: None

  Tag: <clusternode> Per Node configuration
  Parent Tag: <clusternodes>
  Attributes:

  • name(Required): The hostname or IP Address of the node
  • votes(Optional - default is 1): number of votes node can cast
  • nodeid (Required): Each node must have a unique integer value node ID. A node’s application to join the cluster
    may be rejected if you try to set the nodeid to one that
    is already used.
  Example: ==== <clusternode nodeid="3" name="foo.bar.com" votes="3"/> ====

  Tag: <fence> fencing information for the node.
  Parent Tag: <clusternode>
  Attributes: None

  • Tag: <method> defines a method for fencing the machine.
    Parent Tag: <fence>
    Attributes:

    • name (required): name for the fencing method, if GUI generated, it is a number.

  • Tag: <device> The device used to fence the node
    Parent Tag: <method>
    Attributes: See <fencedevice> section of this document for agent specific attributes for the <device> tag.

  Example: ==== <fence> <method name="1"> <device name="apcswitch" port="3" switch="0"/> </method> </fence> ====

  ====

Tag: <rm> The RM block holds resources, failover domains and any number of 'group' (\= resourcegroup) blocks
Parent Tag: <cluster>
Attributes:

• log_level (Optional): An integer 0-7, inclusive for all
  levels less than the selected. 0, system is unusable, emergency;
  1 action must be taken immediately;
  2, critical conditions;
  3, error conditions;
  4, warning conditions;
  5, normal but significant condition;
  6, informational;
  7, debug-level messages.
• log_facility The facility is one of the following keywords:
  auth, authpriv, cron, daemon, kern, lpr, mail, news, syslog,
  user, uucp and local0 through local7.
• Example: ==== <rm log_level="6" log_facility="local4"> </rm> ====

Tag: <failoverdomains> All the <failoverdomain> tags go under here. Parent Tag: <cluster> Attributes: None

Tag: <failoverdomain> Specify properties of specific failover domains Parent Tag: <failoverdomains> Attributes:

• name (Required): the name of the failover domain
• ordered: Set value to '1' if the failover domain is ordered, '0' if not - default is unordered.
• restricted: Set value to '1' if failover domain is restricted, '0' if not - default is unrestricted.

Tag: <failoverdomainnode> An individual node within a failover domain. Parent Tag: <failoverdomain> Attributes:

• name (Required): name of the node.
• priority (Required): a number representing the priority, with lower numbers having higher priority.
Example: ==== <failoverdomains> <failoverdomain name="foo" ordered="1"> <failoverdomainnode name="member" priority="1"/> <failoverdomainnode name="member2" priority="2"/> </failoverdomain> <failoverdomain name="bar" restricted="1"> <failoverdomainnode name="member2" priority="1"/> </failoverdomain> </failoverdomains> ====