|
rpm-build |
3ee90c |
:compat-mode: legacy
|
|
rpm-build |
3ee90c |
= Cluster-Wide Configuration =
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
== Configuration Layout ==
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
The cluster is defined by the Cluster Information Base (CIB),
|
|
rpm-build |
3ee90c |
which uses XML notation. The simplest CIB, an empty one, looks like this:
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
.An empty configuration
|
|
rpm-build |
3ee90c |
======
|
|
rpm-build |
3ee90c |
[source,XML]
|
|
rpm-build |
3ee90c |
-------
|
|
rpm-build |
3ee90c |
<cib crm_feature_set="3.0.7" validate-with="pacemaker-1.2" admin_epoch="1" epoch="0" num_updates="0">
|
|
rpm-build |
3ee90c |
<configuration>
|
|
rpm-build |
3ee90c |
<crm_config/>
|
|
rpm-build |
3ee90c |
<nodes/>
|
|
rpm-build |
3ee90c |
<resources/>
|
|
rpm-build |
3ee90c |
<constraints/>
|
|
rpm-build |
3ee90c |
</configuration>
|
|
rpm-build |
3ee90c |
<status/>
|
|
rpm-build |
3ee90c |
</cib>
|
|
rpm-build |
3ee90c |
-------
|
|
rpm-build |
3ee90c |
======
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
The empty configuration above contains the major sections that make up a CIB:
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
* +cib+: The entire CIB is enclosed with a +cib+ tag. Certain fundamental settings
|
|
rpm-build |
3ee90c |
are defined as attributes of this tag.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
** +configuration+: This section -- the primary focus of this document --
|
|
rpm-build |
3ee90c |
contains traditional configuration information such as what resources the
|
|
rpm-build |
3ee90c |
cluster serves and the relationships among them.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
*** +crm_config+: cluster-wide configuration options
|
|
rpm-build |
3ee90c |
*** +nodes+: the machines that host the cluster
|
|
rpm-build |
3ee90c |
*** +resources+: the services run by the cluster
|
|
rpm-build |
3ee90c |
*** +constraints+: indications of how resources should be placed
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
** +status+: This section contains the history of each resource on each node.
|
|
rpm-build |
3ee90c |
Based on this data, the cluster can construct the complete current
|
|
rpm-build |
3ee90c |
state of the cluster. The authoritative source for this section
|
|
rpm-build |
3ee90c |
is the local executor (pacemaker-execd process) on each cluster node, and
|
|
rpm-build |
3ee90c |
the cluster will occasionally repopulate the entire section. For this
|
|
rpm-build |
3ee90c |
reason, it is never written to disk, and administrators are advised
|
|
rpm-build |
3ee90c |
against modifying it in any way.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
In this document, configuration settings will be described as 'properties' or 'options'
|
|
rpm-build |
3ee90c |
based on how they are defined in the CIB:
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
* Properties are XML attributes of an XML element.
|
|
rpm-build |
3ee90c |
* Options are name-value pairs expressed as +nvpair+ child elements of an XML element.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
Normally, you will use command-line tools that abstract the XML, so the
|
|
rpm-build |
3ee90c |
distinction will be unimportant; both properties and options are
|
|
rpm-build |
3ee90c |
cluster settings you can tweak.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
== CIB Properties ==
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
Certain settings are defined by CIB properties (that is, attributes of the
|
|
rpm-build |
3ee90c |
+cib+ tag) rather than with the rest of the cluster configuration in the
|
|
rpm-build |
3ee90c |
+configuration+ section.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
The reason is simply a matter of parsing. These options are used by the
|
|
rpm-build |
3ee90c |
configuration database which is, by design, mostly ignorant of the content it
|
|
rpm-build |
3ee90c |
holds. So the decision was made to place them in an easy-to-find location.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
.CIB Properties
|
|
rpm-build |
3ee90c |
[width="95%",cols="2m,<5",options="header",align="center"]
|
|
rpm-build |
3ee90c |
|=========================================================
|
|
rpm-build |
3ee90c |
|Field |Description
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| admin_epoch |
|
|
rpm-build |
3ee90c |
indexterm:[Configuration Version,Cluster]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,Configuration Version]
|
|
rpm-build |
3ee90c |
indexterm:[admin_epoch,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,admin_epoch]
|
|
rpm-build |
3ee90c |
When a node joins the cluster, the cluster performs a check to see
|
|
rpm-build |
3ee90c |
which node has the best configuration. It asks the node with the highest
|
|
rpm-build |
3ee90c |
(+admin_epoch+, +epoch+, +num_updates+) tuple to replace the configuration on
|
|
rpm-build |
3ee90c |
all the nodes -- which makes setting them, and setting them correctly, very
|
|
rpm-build |
3ee90c |
important. +admin_epoch+ is never modified by the cluster; you can use this
|
|
rpm-build |
3ee90c |
to make the configurations on any inactive nodes obsolete. _Never set this
|
|
rpm-build |
3ee90c |
value to zero_. In such cases, the cluster cannot tell the difference between
|
|
rpm-build |
3ee90c |
your configuration and the "empty" one used when nothing is found on disk.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| epoch |
|
|
rpm-build |
3ee90c |
indexterm:[epoch,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,epoch]
|
|
rpm-build |
3ee90c |
The cluster increments this every time the configuration is updated (usually by
|
|
rpm-build |
3ee90c |
the administrator).
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| num_updates |
|
|
rpm-build |
3ee90c |
indexterm:[num_updates,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,num_updates]
|
|
rpm-build |
3ee90c |
The cluster increments this every time the configuration or status is updated
|
|
rpm-build |
3ee90c |
(usually by the cluster) and resets it to 0 when epoch changes.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| validate-with |
|
|
rpm-build |
3ee90c |
indexterm:[validate-with,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,validate-with]
|
|
rpm-build |
3ee90c |
Determines the type of XML validation that will be done on the configuration.
|
|
rpm-build |
3ee90c |
If set to +none+, the cluster will not verify that updates conform to the
|
|
rpm-build |
3ee90c |
DTD (nor reject ones that don't). This option can be useful when
|
|
rpm-build |
3ee90c |
operating a mixed-version cluster during an upgrade.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
|cib-last-written |
|
|
rpm-build |
3ee90c |
indexterm:[cib-last-written,Cluster Property]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Property,cib-last-written]
|
|
rpm-build |
3ee90c |
Indicates when the configuration was last written to disk. Maintained by the
|
|
rpm-build |
3ee90c |
cluster; for informational purposes only.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
|have-quorum |
|
|
rpm-build |
3ee90c |
indexterm:[have-quorum,Cluster Property]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Property,have-quorum]
|
|
rpm-build |
3ee90c |
Indicates if the cluster has quorum. If false, this may mean that the
|
|
rpm-build |
3ee90c |
cluster cannot start resources or fence other nodes (see
|
|
rpm-build |
3ee90c |
+no-quorum-policy+ below). Maintained by the cluster.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
|dc-uuid |
|
|
rpm-build |
3ee90c |
indexterm:[dc-uuid,Cluster Property]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Property,dc-uuid]
|
|
rpm-build |
3ee90c |
Indicates which cluster node is the current leader. Used by the
|
|
rpm-build |
3ee90c |
cluster when placing resources and determining the order of some
|
|
rpm-build |
3ee90c |
events. Maintained by the cluster.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
|=========================================================
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
[[s-cluster-options]]
|
|
rpm-build |
3ee90c |
== Cluster Options ==
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
Cluster options, as you might expect, control how the cluster behaves
|
|
rpm-build |
3ee90c |
when confronted with certain situations.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
They are grouped into sets within the +crm_config+ section, and, in advanced
|
|
rpm-build |
3ee90c |
configurations, there may be more than one set. (This will be described later
|
|
rpm-build |
3ee90c |
in the section on <<ch-rules>> where we will show how to have the cluster use
|
|
rpm-build |
3ee90c |
different sets of options during working hours than during weekends.) For now,
|
|
rpm-build |
3ee90c |
we will describe the simple case where each option is present at most once.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
You can obtain an up-to-date list of cluster options, including
|
|
rpm-build |
3ee90c |
their default values, by running the `man pacemaker-schedulerd` and
|
|
rpm-build |
3ee90c |
`man pacemaker-controld` commands.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
.Cluster Options
|
|
rpm-build |
3ee90c |
[width="95%",cols="5m,2,<11",options="header",align="center"]
|
|
rpm-build |
3ee90c |
|=========================================================
|
|
rpm-build |
3ee90c |
|Option |Default |Description
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| cluster-name | |
|
|
rpm-build |
3ee90c |
indexterm:[cluster-name,Cluster Property]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Property,cluster-name]
|
|
rpm-build |
3ee90c |
An (optional) name for the cluster as a whole. This is mostly for users'
|
|
rpm-build |
3ee90c |
convenience for use as desired in administration, but this can be used
|
|
rpm-build |
3ee90c |
in the Pacemaker configuration in <<ch-rules,rules>> (as the
|
|
rpm-build |
3ee90c |
+#cluster-name+ <<node-attribute-expressions-special,node attribute>>). It may
|
|
rpm-build |
3ee90c |
also be used by higher-level tools when displaying cluster information, and by
|
|
rpm-build |
3ee90c |
certain resource agents (for example, the +ocf:heartbeat:GFS2+ agent stores the
|
|
rpm-build |
3ee90c |
cluster name in filesystem meta-data).
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| dc-version | |
|
|
rpm-build |
3ee90c |
indexterm:[dc-version,Cluster Property]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Property,dc-version]
|
|
rpm-build |
3ee90c |
Version of Pacemaker on the cluster's DC.
|
|
rpm-build |
3ee90c |
Determined automatically by the cluster.
|
|
rpm-build |
3ee90c |
Often includes the hash which identifies the exact Git changeset it was built
|
|
rpm-build |
3ee90c |
from. Used for diagnostic purposes.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| cluster-infrastructure | |
|
|
rpm-build |
3ee90c |
indexterm:[cluster-infrastructure,Cluster Property]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Property,cluster-infrastructure]
|
|
rpm-build |
3ee90c |
The messaging stack on which Pacemaker is currently running.
|
|
rpm-build |
3ee90c |
Determined automatically by the cluster.
|
|
rpm-build |
3ee90c |
Used for informational and diagnostic purposes.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| no-quorum-policy | stop
|
|
rpm-build |
3ee90c |
a|
|
|
rpm-build |
3ee90c |
indexterm:[no-quorum-policy,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,no-quorum-policy]
|
|
rpm-build |
3ee90c |
What to do when the cluster does not have quorum. Allowed values:
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
* +ignore:+ continue all resource management
|
|
rpm-build |
3ee90c |
* +freeze:+ continue resource management, but don't recover resources from nodes not in the affected partition
|
|
rpm-build |
3ee90c |
* +stop:+ stop all resources in the affected cluster partition
|
|
rpm-build |
3ee90c |
* +suicide:+ fence all nodes in the affected cluster partition
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| batch-limit | 0 |
|
|
rpm-build |
3ee90c |
indexterm:[batch-limit,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,batch-limit]
|
|
rpm-build |
3ee90c |
The maximum number of actions that the cluster may execute in parallel across
|
|
rpm-build |
3ee90c |
all nodes. The "correct" value will depend on the speed and load of your
|
|
rpm-build |
3ee90c |
network and cluster nodes. If zero, the cluster will impose a dynamically
|
|
rpm-build |
3ee90c |
calculated limit only when any node has high load.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| migration-limit | -1 |
|
|
rpm-build |
3ee90c |
indexterm:[migration-limit,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,migration-limit]
|
|
rpm-build |
3ee90c |
The number of <<s-migrating-resources,live migration>> actions that the cluster
|
|
rpm-build |
3ee90c |
is allowed to execute in parallel on a node. A value of -1 means unlimited.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| symmetric-cluster | TRUE |
|
|
rpm-build |
3ee90c |
indexterm:[symmetric-cluster,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,symmetric-cluster]
|
|
rpm-build |
3ee90c |
Can all resources run on any node by default?
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| stop-all-resources | FALSE |
|
|
rpm-build |
3ee90c |
indexterm:[stop-all-resources,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,stop-all-resources]
|
|
rpm-build |
3ee90c |
Should the cluster stop all resources?
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| stop-orphan-resources | TRUE |
|
|
rpm-build |
3ee90c |
indexterm:[stop-orphan-resources,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,stop-orphan-resources]
|
|
rpm-build |
3ee90c |
Should deleted resources be stopped? This value takes precedence over
|
|
rpm-build |
3ee90c |
+is-managed+ (i.e. even unmanaged resources will be stopped if deleted from
|
|
rpm-build |
3ee90c |
the configuration when this value is TRUE).
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| stop-orphan-actions | TRUE |
|
|
rpm-build |
3ee90c |
indexterm:[stop-orphan-actions,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,stop-orphan-actions]
|
|
rpm-build |
3ee90c |
Should deleted actions be cancelled?
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| start-failure-is-fatal | TRUE |
|
|
rpm-build |
3ee90c |
indexterm:[start-failure-is-fatal,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,start-failure-is-fatal]
|
|
rpm-build |
3ee90c |
Should a failure to start a resource on a particular node prevent further start
|
|
rpm-build |
3ee90c |
attempts on that node? If FALSE, the cluster will decide whether the same
|
|
rpm-build |
3ee90c |
node is still eligible based on the resource's current failure count
|
|
rpm-build |
3ee90c |
and +migration-threshold+ (see <<s-failure-handling>>).
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| enable-startup-probes | TRUE |
|
|
rpm-build |
3ee90c |
indexterm:[enable-startup-probes,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,enable-startup-probes]
|
|
rpm-build |
3ee90c |
Should the cluster check for active resources during startup?
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| maintenance-mode | FALSE |
|
|
rpm-build |
3ee90c |
indexterm:[maintenance-mode,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,maintenance-mode]
|
|
rpm-build |
3ee90c |
Should the cluster refrain from monitoring, starting and stopping resources?
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| stonith-enabled | TRUE |
|
|
rpm-build |
3ee90c |
indexterm:[stonith-enabled,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,stonith-enabled]
|
|
rpm-build |
3ee90c |
Should failed nodes and nodes with resources that can't be stopped be
|
|
rpm-build |
3ee90c |
shot? If you value your data, set up a STONITH device and enable this.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
If true, or unset, the cluster will refuse to start resources unless
|
|
rpm-build |
3ee90c |
one or more STONITH resources have been configured.
|
|
rpm-build |
3ee90c |
If false, unresponsive nodes are immediately assumed to be running no
|
|
rpm-build |
3ee90c |
resources, and resource takeover to online nodes starts without any
|
|
rpm-build |
3ee90c |
further protection (which means _data loss_ if the unresponsive node
|
|
rpm-build |
3ee90c |
still accesses shared storage, for example). See also the +requires+
|
|
rpm-build |
3ee90c |
meta-attribute in <<s-resource-options>>.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| stonith-action | reboot |
|
|
rpm-build |
3ee90c |
indexterm:[stonith-action,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,stonith-action]
|
|
rpm-build |
3ee90c |
Action to send to STONITH device. Allowed values are +reboot+ and +off+.
|
|
rpm-build |
3ee90c |
The value +poweroff+ is also allowed, but is only used for
|
|
rpm-build |
3ee90c |
legacy devices.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| stonith-timeout | 60s |
|
|
rpm-build |
3ee90c |
indexterm:[stonith-timeout,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,stonith-timeout]
|
|
rpm-build |
3ee90c |
How long to wait for STONITH actions (reboot, on, off) to complete
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| stonith-max-attempts | 10 |
|
|
rpm-build |
3ee90c |
indexterm:[stonith-max-attempts,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,stonith-max-attempts]
|
|
rpm-build |
3ee90c |
How many times fencing can fail for a target before the cluster will no longer
|
|
rpm-build |
3ee90c |
immediately re-attempt it.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| stonith-watchdog-timeout | 0 |
|
|
rpm-build |
3ee90c |
indexterm:[stonith-watchdog-timeout,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,stonith-watchdog-timeout]
|
|
rpm-build |
3ee90c |
If nonzero, rely on hardware watchdog self-fencing. If positive, assume unseen
|
|
rpm-build |
3ee90c |
nodes self-fence within this much time. If negative, and the
|
|
rpm-build |
3ee90c |
SBD_WATCHDOG_TIMEOUT environment variable is set, use twice that value.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| concurrent-fencing | FALSE |
|
|
rpm-build |
3ee90c |
indexterm:[concurrent-fencing,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,concurrent-fencing]
|
|
rpm-build |
3ee90c |
Is the cluster allowed to initiate multiple fence actions concurrently?
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| fence-reaction | stop |
|
|
rpm-build |
3ee90c |
indexterm:[fence-reaction,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,fence-reaction]
|
|
rpm-build |
3ee90c |
How should a cluster node react if notified of its own fencing? A cluster node
|
|
rpm-build |
3ee90c |
may receive notification of its own fencing if fencing is misconfigured, or if
|
|
rpm-build |
3ee90c |
fabric fencing is in use that doesn't cut cluster communication. Allowed values
|
|
rpm-build |
3ee90c |
are +stop+ to attempt to immediately stop pacemaker and stay stopped, or
|
|
rpm-build |
3ee90c |
+panic+ to attempt to immediately reboot the local node, falling back to stop
|
|
rpm-build |
3ee90c |
on failure. The default is likely to be changed to +panic+ in a future release.
|
|
rpm-build |
3ee90c |
'(since 2.0.3)'
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| cluster-delay | 60s |
|
|
rpm-build |
3ee90c |
indexterm:[cluster-delay,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,cluster-delay]
|
|
rpm-build |
3ee90c |
Estimated maximum round-trip delay over the network (excluding action
|
|
rpm-build |
3ee90c |
execution). If the DC requires an action to be executed on another
|
|
rpm-build |
3ee90c |
node, it will consider the action failed if it does not get a response
|
|
rpm-build |
3ee90c |
from the other node in this time (after considering the action's
|
|
rpm-build |
3ee90c |
own timeout). The "correct" value will depend on the speed and load of your
|
|
rpm-build |
3ee90c |
network and cluster nodes.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| dc-deadtime | 20s |
|
|
rpm-build |
3ee90c |
indexterm:[dc-deadtime,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,dc-deadtime]
|
|
rpm-build |
3ee90c |
How long to wait for a response from other nodes during startup.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
The "correct" value will depend on the speed/load of your network and the type of switches used.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| cluster-ipc-limit | 500 |
|
|
rpm-build |
3ee90c |
indexterm:[cluster-ipc-limit,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,cluster-ipc-limit]
|
|
rpm-build |
3ee90c |
The maximum IPC message backlog before one cluster daemon will disconnect
|
|
rpm-build |
3ee90c |
another. This is of use in large clusters, for which a good value is the number
|
|
rpm-build |
3ee90c |
of resources in the cluster multiplied by the number of nodes. The default of
|
|
rpm-build |
3ee90c |
500 is also the minimum. Raise this if you see "Evicting client" messages for
|
|
rpm-build |
3ee90c |
cluster daemon PIDs in the logs.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| pe-error-series-max | -1 |
|
|
rpm-build |
3ee90c |
indexterm:[pe-error-series-max,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,pe-error-series-max]
|
|
rpm-build |
3ee90c |
The number of PE inputs resulting in ERRORs to save. Used when reporting problems.
|
|
rpm-build |
3ee90c |
A value of -1 means unlimited (report all).
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| pe-warn-series-max | -1 |
|
|
rpm-build |
3ee90c |
indexterm:[pe-warn-series-max,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,pe-warn-series-max]
|
|
rpm-build |
3ee90c |
The number of PE inputs resulting in WARNINGs to save. Used when reporting problems.
|
|
rpm-build |
3ee90c |
A value of -1 means unlimited (report all).
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| pe-input-series-max | -1 |
|
|
rpm-build |
3ee90c |
indexterm:[pe-input-series-max,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,pe-input-series-max]
|
|
rpm-build |
3ee90c |
The number of "normal" PE inputs to save. Used when reporting problems.
|
|
rpm-build |
3ee90c |
A value of -1 means unlimited (report all).
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| placement-strategy | default |
|
|
rpm-build |
3ee90c |
indexterm:[placement-strategy,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,placement-strategy]
|
|
rpm-build |
3ee90c |
How the cluster should allocate resources to nodes (see <<s-utilization>>).
|
|
rpm-build |
3ee90c |
Allowed values are +default+, +utilization+, +balanced+, and +minimal+.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| node-health-strategy | none |
|
|
rpm-build |
3ee90c |
indexterm:[node-health-strategy,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,node-health-strategy]
|
|
rpm-build |
3ee90c |
How the cluster should react to node health attributes (see <<s-node-health>>).
|
|
rpm-build |
3ee90c |
Allowed values are +none+, +migrate-on-red+, +only-green+, +progressive+, and
|
|
rpm-build |
3ee90c |
+custom+.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| enable-acl | FALSE |
|
|
rpm-build |
3ee90c |
indexterm:[enable-acl,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,enable-acl]
|
|
rpm-build |
3ee90c |
Whether access control lists (ACLs) (see <<ch-acls>>) can be used to authorize
|
|
rpm-build |
3ee90c |
modifications to the CIB.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| node-health-base | 0 |
|
|
rpm-build |
3ee90c |
indexterm:[node-health-base,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,node-health-base]
|
|
rpm-build |
3ee90c |
The base health score assigned to a node. Only used when
|
|
rpm-build |
3ee90c |
+node-health-strategy+ is +progressive+.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| node-health-green | 0 |
|
|
rpm-build |
3ee90c |
indexterm:[node-health-green,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,node-health-green]
|
|
rpm-build |
3ee90c |
The score to use for a node health attribute whose value is +green+.
|
|
rpm-build |
3ee90c |
Only used when +node-health-strategy+ is +progressive+ or +custom+.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| node-health-yellow | 0 |
|
|
rpm-build |
3ee90c |
indexterm:[node-health-yellow,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,node-health-yellow]
|
|
rpm-build |
3ee90c |
The score to use for a node health attribute whose value is +yellow+.
|
|
rpm-build |
3ee90c |
Only used when +node-health-strategy+ is +progressive+ or +custom+.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| node-health-red | 0 |
|
|
rpm-build |
3ee90c |
indexterm:[node-health-red,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,node-health-red]
|
|
rpm-build |
3ee90c |
The score to use for a node health attribute whose value is +red+.
|
|
rpm-build |
3ee90c |
Only used when +node-health-strategy+ is +progressive+ or +custom+.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| cluster-recheck-interval | 15min |
|
|
rpm-build |
3ee90c |
indexterm:[cluster-recheck-interval,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,cluster-recheck-interval]
|
|
rpm-build |
3ee90c |
Pacemaker is primarily event-driven, and looks ahead to know when to recheck
|
|
rpm-build |
3ee90c |
the cluster for failure timeouts and most time-based rules. However, it will
|
|
rpm-build |
3ee90c |
also recheck the cluster after this amount of inactivity. This has two goals:
|
|
rpm-build |
3ee90c |
rules with +date_spec+ are only guaranteed to be checked this often, and it
|
|
rpm-build |
3ee90c |
also serves as a fail-safe for certain classes of scheduler bugs. A value of 0
|
|
rpm-build |
3ee90c |
disables this polling; positive values are a time interval.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| remove-after-stop | FALSE |
|
|
rpm-build |
3ee90c |
indexterm:[remove-after-stop,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,remove-after-stop]
|
|
rpm-build |
3ee90c |
_Advanced Use Only:_ Should the cluster remove resources from the LRM after
|
|
rpm-build |
3ee90c |
they are stopped? Values other than the default are, at best, poorly tested and
|
|
rpm-build |
3ee90c |
potentially dangerous.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| startup-fencing | TRUE |
|
|
rpm-build |
3ee90c |
indexterm:[startup-fencing,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,startup-fencing]
|
|
rpm-build |
3ee90c |
_Advanced Use Only:_ Should the cluster shoot unseen nodes?
|
|
rpm-build |
3ee90c |
Not using the default is very unsafe!
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| election-timeout | 2min |
|
|
rpm-build |
3ee90c |
indexterm:[election-timeout,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,election-timeout]
|
|
rpm-build |
3ee90c |
_Advanced Use Only:_ If you need to adjust this value, it probably indicates
|
|
rpm-build |
3ee90c |
the presence of a bug.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| shutdown-escalation | 20min |
|
|
rpm-build |
3ee90c |
indexterm:[shutdown-escalation,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,shutdown-escalation]
|
|
rpm-build |
3ee90c |
_Advanced Use Only:_ If you need to adjust this value, it probably indicates
|
|
rpm-build |
3ee90c |
the presence of a bug.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| join-integration-timeout | 3min |
|
|
rpm-build |
3ee90c |
indexterm:[join-integration-timeout,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,join-integration-timeout]
|
|
rpm-build |
3ee90c |
_Advanced Use Only:_ If you need to adjust this value, it probably indicates
|
|
rpm-build |
3ee90c |
the presence of a bug.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| join-finalization-timeout | 30min |
|
|
rpm-build |
3ee90c |
indexterm:[join-finalization-timeout,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,join-finalization-timeout]
|
|
rpm-build |
3ee90c |
_Advanced Use Only:_ If you need to adjust this value, it probably indicates
|
|
rpm-build |
3ee90c |
the presence of a bug.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
| transition-delay | 0s |
|
|
rpm-build |
3ee90c |
indexterm:[transition-delay,Cluster Option]
|
|
rpm-build |
3ee90c |
indexterm:[Cluster,Option,transition-delay]
|
|
rpm-build |
3ee90c |
_Advanced Use Only:_ Delay cluster recovery for the configured interval to
|
|
rpm-build |
3ee90c |
allow for additional/related events to occur. Useful if your configuration is
|
|
rpm-build |
3ee90c |
sensitive to the order in which ping updates arrive.
|
|
rpm-build |
3ee90c |
Enabling this option will slow down cluster recovery under
|
|
rpm-build |
3ee90c |
all conditions.
|
|
rpm-build |
3ee90c |
|
|
rpm-build |
3ee90c |
|=========================================================
|