HAGRP(1M)

NAME

hagrp - perform VCS service group operations

SYNOPSIS

hagrp -add group

hagrp -delete group

hagrp -link parent_group child_group dependency

hagrp -local group attr

hagrp -global group attr                value           ... | key ... | { key value }

hagrp -unlink parent_group child_group

hagrp -clear group [-sys system] [-clus cluster | -localclus]

hagrp -clearadminwait [-fault] group -sys system

hagrp -flush [-force] group -sys system [-clus cluster | -localclus]

hagrp -forecast group [ -verbose]

hagrp -online [-force] [-nopre] group -sys system [-checkpartial system] [-clus cluster | -localclus]

hagrp -online [-force] group -any [-clus cluster | -localclus]

hagrp -online group -site site

hagrp -online [-propagate] group -sys system

hagrp -online [-propagate] group -any

hagrp -online [-propagate] group -site site

hagrp -offline [-force] [-ifprobed] group -sys system [-clus cluster | -localclus]

hagrp -offline [-force] [-ifprobed] group -any [-clus cluster | -localclus]

hagrp -offline group -site site

hagrp -offline [-propagate] group -sys system

hagrp -offline [-propagate] group -any

hagrp -offline [-propagate] group -site site

hagrp -switch group -to system [-clus cluster | -localclus [-nopre]]

hagrp -switch group -any [-clus cluster | -localclus]

hagrp -switch group -site site

hagrp -freeze group -persistent

hagrp -freeze group [-clus cluster | -localclus]

hagrp -unfreeze group -persistent

hagrp -unfreeze group [-clus cluster | -localclus]

hagrp -enable group [-sys system]

hagrp -disable group [-sys system]

hagrp -autoenable group -sys system

hagrp -enableresources group

hagrp -disableresources group

hagrp -ignore -group -clus cluster

hagrp -display [group(s)] [-attribute attribute(s)] [-sys system(s)] [-clus cluster | -localclus]

hagrp -state [group(s)] [-sys system(s)] [-clus cluster | -localclus]

hagrp -list [conditional(s)] [-clus cluster | -localclus]

hagrp -value group attr [system] [-clus cluster | -localclus]

hagrp -dep [group(s)]

hagrp -resources group [-clus cluster | -localclus]

hagrp -wait group attribute value[-clus cluster] [-sys system] [-time timeout]

hagrp -help [-modify | -link | -list]

hagrp -modify modify_options

hagrp -migrate service_group -to system

AVAILABILITY

VRTSvcs

DESCRIPTION

A service group is an instance of a computer service that is made highly available with VCS. A group comprises one or more resources of various resource types, such as disks, volumes, or databases. Use the hagrp command to view information about and manage each group. See NOTES for how to specify "-" and "%" characters in the command line.

OPTIONS

-add group
  Add a service group to the cluster.
-delete group
  Delete a service group. The group must have no resources. To delete resources, use the hares command.
-link parent_group child_group dependency
  Specify dependencies between service groups. The variable dependency is one of the following:
online global [ firm | soft ]
  The child group must be online on some system in the cluster before the parent group can come online. With the dependency set to firm, if the child group faults, the parent group must go offline until the child group fails over to another system, at which time the parent can return online. If VCS cannot bring the child group online in the cluster, the parent group remains offline. With the dependency set to soft, if the child group faults and fails over, the parent group continues online. If VCS cannot bring the child group online in the cluster, the parent group remains online.
online local [ firm | soft | hard ]
  The child group must be online on the same system in the cluster before the parent group can go online. With the dependency set to firm, if the child group faults, the parent group must go offline. If the child fails over, the parent group comes back online on the same system as the child. If VCS cannot bring the child group online in the cluster, the parent group remains offline. With the dependency set to soft, if the child group faults, the parent group continues to run on the local system until the child fails over to another system in the cluster, at which time the parent group will fail over to the same system as the child. If VCS cannot bring the child group online in the cluster, the parent group remains online. With the dependency set to hard, if the child group faults, the parent group must go offline. If the child fails over, the parent group comes back online on the system as the child. If the parent group faults, the child group must go offline. If the child fails over, parent group comes back online on the same system as the child.
online remote [ firm | soft ]
  The child must be online on any other system in the cluster before the parent can go online. With the dependency set to firm, if the child group faults, the parent group goes offline. When the child group fails over to another system in the cluster, the parent group comes back online on a system other than the system on which the child is online. If VCS cannot bring the child group online in the cluster, the parent group remains offline. With the dependency set to soft, if the child group faults, the parent group continues to remain online unless VCS attempts to fail over the child group to the system on which the parent group is running; in this case, the parent group is taken offline, to be brought online elsewhere. If the child group cannot be brought online anywhere in the cluster, the parent stays online.
online site [ firm | soft ]
  The child group must be online on any system associated to the same site in the cluster before the parent group can go online. With the dependency set to firm, if the child group faults, the parent group must go offline. If the child fails over, the parent group comes back online on the same site as the child. If VCS cannot bring the child group online in the cluster, the parent group remains offline. With the dependency set to soft, if the child group fails over to another site then the parent group fails over to the same site as the child. However, if child group fails over to another system in the same site then the parent group remains online on the same site.
offline local
  The child group must be offline on the same system before the parent group can go online. Similarly the parent group must be offline on the same system before the child group can go online.
A group dependency tree may be at most five levels deep, and parent groups with dependencies except offline local and online local hard can have more than one child. Parallel parent groups dependent on parallel child groups are not supported in online global or online remote dependencies. The configuration of parallel parent groups that are dependent on a failover child group is not supported in online local dependencies.
-local group attr
  Localize an attribute of a service group. The command changes the scope of a global attribute to a local attribute. That is, the current attribute value is converted to an association in which the keys are the systems of the group\(cqs SystemList attribute. The PreOnline, ContainerInfo, TriggersEnabled , and UserAssoc attributes can be localized using the command. Localized attributes may have a different value for each system in the SystemList.
-global group attr value ... | key ... | { key value}
  Change the scope of a local attribute (one that has a value or set of values for every system on which a group is configured to run) to the scope of a global attribute (a single value or a set of values for all systems). This value is the one specified in the command. An attribute cannot have both local and global values defined in a service group. For example, the PreOnline attribute cannot be defined with both local and global scope. The PreOnline, ContainerInfo, TriggersEnabled, and UserAssoc attributes can be made global using the command.
-unlink parent_group child_group
  Remove dependency between two service groups. Note that the dependency is not specified.
-clear group [-sys system] [-clus cluster | -localclus]
  Clear all faulted, non-persistent resources in the specified group by changing their state from faulted to offline. If no system is specified, all resources are cleared on all systems in the group’s SystemList. A message is printed if no faulted, non-persistent resources exist. If the Global Cluster option is in use, the command clears non-persistent faulted resources for the specified system or the specified cluster. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. For a system whose name is not unique among clusters, specify the name of its cluster. Only global service groups can be cleared from another cluster.
-clearadminwait [-fault] group -sys system
  Clear the ADMIN_WAIT state of all resources in the specified group on the specified system. If the resources continue in the ADMIN_WAIT state, use the -fault option to clear the ADMIN_WAIT state. The state is set to ONLINE | UNABLE_TO_OFFLINE or FAULTED, depending on the reasons the ResAdminWait trigger had been called. Note that the online, offline, switch, and flush operations cannot be performed on resources in the ADMIN_WAIT state. Also, when resources are in the ADMIN_WAIT state, the hastop command requires the -force option.
-flush [-force] group -sys system [-clus cluster | -localclus]
  Flush a service group and enable corrective action. All resources in the group that are waiting to go online automatically transition to not waiting. Resources in the waiting to go offline state remain in that state. For resources in the waiting to go offline state, the -force option can be used to transition resources to the not waiting state. Failovers and switches in progress are canceled. If the Global Cluster option is in use, the command flushes global service groups for corrective action on the specified system, local or remote. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. For a system whose name is not unique among clusters, specify the name of its cluster. Only global service groups can be flushed from another cluster.
-forecast group [ -verbose]
  For a service group without virtual machine resources, the -forecast option displays the online node, the target node, and the FailOverPolicy value that is specified either with the policy option or the groups FailOverPolicy attribute for the service group. Note that the target node is evaluated based on the HostAvailableForecast attribute values.

For a service group with virtual machine resources, the -forecast option for the target node is based on the ServerAvailableForecast attribute values.

The -verbose option displays the online node, the target node, and the FailOverPolicy attribute value for the service group. It also displays the MeterWeight and Load attribute values. The BiggestAvailable policy considers the MeterWeight and Load attributes when selecting a target node.

If the MeterWeight attribute is not configured at the group level, its value is retrieved from the Cluster::Meterweight attribute. The -verbose option also displays the systems that are considered during the target node selection. The Forecasted Free/Total Capacity displays the available and total capacity of the system on whether the system met the load requirements for the group. The System weight attribute is a number that is computed based on the available capacity of the system and the MeterWeight attribute. The command also marks the target system with an asterisk (*).

-online [-force] [-nopre] group -sys system [-checkpartial system]
  [-clus cluster | -localclus]
Start a service group by bringing its resources online on a specified system. Resources that have their AutoStart attribute set to zero (the default is one) are not started by this command unless resources that have AutoStart set to one depend on the resources. The -nopre option can be used within a user-defined Preonline event trigger script to bring a service group online without executing the Preonline trigger again. The -checkpartial option is used by the Preonline trigger during failover. When a service group configured with Preonline=1 on one system (system 1) fails over to another system (system 2), only those resources online on system 1 prior to failover are brought online on system 2. If the Global Cluster option is in use, the command brings the specified global service group online on the specified system in the specified cluster. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. The -force option can be used to online a global group on a cluster when another cluster having the authority for the group is faulted. For a system whose name is not unique among clusters, specify the name of its cluster. Only global service groups can be brought online from another cluster.
-online [-force] group -any [-clus cluster | -localclus]
  The -any option can be used to start a service group by bringing its resources online on the best possible system in a cluster, based on the value of the group’s FailOverPolicy attribute. A parallel service group is brought online on multiple systems in a cluster if so configured. A hybrid service group (group with Parallel attribute set to 2) is brought online on any system in each site. Resources that have their AutoStart attribute set to zero (the default is one) are not started by this command unless resources that have AutoStart set to one depend on the resources. If the Global Cluster option is in use, the command brings the specified global service group online on any system in the specified cluster. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. The -force option can be used to online a global group on a cluster when another cluster having the authority for the group is faulted. Only global service groups can be brought online from another cluster.
-online group -site site
  The -site option can be used to start a service group on any system in the site by bringing its resources online. Resources that have their AutoStart attribute set to zero (the default is one) are not started by this command unless resources that have AutoStart set to one depend on the resources. If the service group is parallel (group with Parallel attribute set to 1), then the service group is brought online on all systems in the site.
-online [-propagate] group -sys system
  The -propagate option specifies that, when you bring a group online, all of its required child groups are also brought online. Resources that have their AutoStart attribute set to zero (the default is one) are not started by this command unless resources that have AutoStart set to one depend on the resources. For example, if G1 depends on G2, which depends on G3, when G1 is brought online with the -propagate option, G2 and G3 are brought online, if they are not online. The -propagate option applies for all child groups, including those with local hard, local firm, local soft, global firm, global soft, remote firm, remote soft, site soft, and site firm dependencies. Note that the specified service group must not be currently in the process of coming online, going offline, or failing over to another system.
-online [-propagate] group -any
  The -any option can be used to bring a service group and all of its required child groups online, on the best possible system in a cluster, based on the value of the group’s FailOverPolicy attribute. A parallel service group is brought online on multiple systems in a cluster if so configured. Resources that have their AutoStart attribute set to zero (the default is one) are not started by this command unless resources that have AutoStart set to one depend on the resources. For example, if G1 depends on G2, which depends on G3, when G1 is brought online with the -propagate option, G2 and G3 are brought online if they are not online. The -propagate option applies for all child groups, including those with local hard, local firm, local soft, global firm, global soft, remote firm, remote soft, site firm and site soft dependencies. Note that the specified service group must not be currently in the process of coming online, going offline, or failing over to another system.
-online [-propagate] group -site site
  The -propagate option specifies that you bring a group and all its child groups online. The parent group is brought online on any system in the site. However, child groups are brought online on a system based on their dependency type. Child groups with global dependency type are brought online on any system in the cluster, while child groups with site dependency are brought online on any system in the same site. Resources that have their AutoStart attribute set to zero (the default is one) are not started by this command unless resources that have AutoStart set to one depend on the resources. For example, if G1 depends on G2, which depends on G3, when G1 is brought online with the -propagate option, G2 and G3 are brought online, if they are not online. The -propagate option applies for all child groups, including those with local hard, local firm, local soft, global firm, global soft, remote firm, remote soft, site firm, and site soft dependencies.
-offline [-force] [-ifprobed] group -sys system [-clus cluster |
  -localclus]
Stop a service group by taking its resources offline on the specified system. The -force option is required to offline the ClusterService group.
Note: it is not recommended to offline the GlobalService group when the Global Cluster option is in use. If the Global Cluster option is in use, the command takes the specified global service group offline on specified system in the specified cluster. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. For a system whose name is not unique among clusters, specify the name of its cluster. Only global service groups can be taken offline from another cluster.
-offline [-force] [-ifprobed] group -any [-clus cluster | -localclus]
  The -any option can be used to stop a service group by taking its resources offline on any system within a cluster. The -force option is required to offline the ClusterService group.
Note: Veritas recommends that you do not offline the GlobalService group when you use global clusters. If you use global clusters, the command takes the specified global service group offline on any system in the specified cluster. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. The -force option can be used to offline a global group on a cluster when another cluster having the authority for the group is faulted. Only global service groups can be taken offline from another cluster.
-offline group -site site
  Stop the specified service group on all the systems in the site.
-offline [-propagate] group -sys system
  The -propagate option specifies that, when you bring a group offline, all of its required parent groups with hard and firm dependencies are also brought offline. Parents with a soft dependency are not taken offline, because parents with a soft dependency can remain online even when the child group is brought offline. For example, if G1 has a firm dependency on G2, and G2 has a firm dependency on G3, then when the command to offline G3 is issued with the -propagate option, G1, G2 and G3 are taken offline.

The -propagate option applies for all parent groups, including those with local hard, local firm, local soft, global firm, global soft, remote firm, and remote soft dependencies.

-offline [-propagate] group -any
  The -any option can be used to bring a group and all of its required parent groups with hard and firm dependencies offline, on any system within a cluster. Parents with a soft dependency are not taken offline, because parents with a soft dependency can remain online even when the child group is brought offline. For example, if G1 has a firm dependency on G2, and G2 has a firm dependency on G3, then when the command to offline G3 is issued with the -propagate option, G1, G2 and G3 are taken offline.

The -propagate option applies for all parent groups, including those with local hard, local firm, local soft, global firm, global soft, remote firm, and remote soft dependencies.

-offline [-propagate] group -site site
  The -propagate option specifies that you take a group offline and all of its required parent groups are taken offline. The group is taken offline on all the systems in the site. However, parent groups are taken offline based on their dependency type. Parent groups with global dependency are taken offline on all the systems in the cluster, while parent groups with site dependency are taken offline on all systems in the site. Parents with a soft dependency are not taken offline, because parents with a soft dependency can remain online even when the child group is brought offline. For example, if G1 has a firm dependency on G2, and G2 has a firm dependency on G3, then when the command to offline G3 is issued with the -propagate option, G1, G2 and G3 are taken offline. The -propagate option applies for all parent groups, including those with local hard, local firm, local soft, global firm, global soft, remote firm, and remote soft, site firm, and site soft dependencies.
-switch group -to system [-clus cluster | -localclus [-nopre]]
  Switch a service group from the system on which it is active to the specified system. This applies only to failover groups (groups that have the Parallel attribute set to zero).

If you use global clusters, the command switches a global service group from a system to a specified system in the specified remote cluster. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. Only global service groups can be switched within or across clusters from another cluster. If you set the value of the PreSwitch group attribute to 1 in the global group definition of the remote cluster, the VCS engine invokes PreSwitch action for the resources that support the PreSwitch action. The engine switches a service group normally if all the actions succeed. If any of the actions fail, then the engine aborts the switch operation. See Cluster Server Administrator’s Guide for more information.

The -nopre option indicates that the VCS engine must switch the service group regardless of the value of the PreSwitch service group attribute.

-switch group -any [-clus cluster | -localclus]
  The -any option can be used to switch a service group to the best possible system on which it is currently not online, based on the value of the group’s FailOverPolicy attribute.

If you use global clusters, the command switches a global service group from a system to a system in specified cluster, either local or remote. The -clus or -localclus option indicates a specific cluster. The -localclus option specifies the cluster in which the command is issued. Only global service groups can be switched within or across clusters from another cluster. With the -any option, the VCS engine does not invoke the PreSwitch action of any resources even if you set the value of the PreSwitch group attribute to 1.

If you run this command to switch a parallel global service group across clusters, VCS brings the parallel service group online on all possible nodes in the remote cluster.

-switch group -site site
  The -switch option can be used to switch a service group from the system on which it is active to any system in the site. The switch operation applies only to failover groups (groups that have the Parallel attribute set to zero).
-freeze group -persistent
  Freeze a service group (disables onlining, offlining, and failover). The -persistent flag indicates the freeze is maintained after a system is rebooted, or the engine is stopped and restarted. NOTE that the configuration must be read-write to use this option.
-freeze group [-clus cluster | -localclus]
  Freeze a service group (disable onlining, offlining, and failover). If the Global Cluster option is in use, the command freezes the specified global service group in the specified cluster. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. Only global service groups can be frozen from another cluster.
-unfreeze group -persistent
  Thaw a service group (reenables onlining, offlining, and failover). The -persistent flag indicates that unfreeze is maintained after a system is rebooted or after the VCS engine is stopped and restarted. Note that the configuration must be read-write to use this option.
-unfreeze group [-clus cluster | -localclus]
  Thaw a service group (reenables onlining, offlining, and failover). If the Global Cluster option is in use, the command unfreezes the specified global service group in the specified cluster. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. Only global service groups can be unfrozen from another cluster.
-enable group [-sys system]
  Enable a service group.
-disable group [-sys system]
  Disable a service group. Actions such as onlining and switching are not permitted on the group.
-autoenable group -sys system
  When VCS does not know the status of a given service group on a given system, it marks that service group as auto-disabled by setting the AutoDisabled attribute to 1 for that system. This can happen if VCS is not running on that system, or if all the resources in the service group on that system are not probed yet. A service group cannot be onlined or failed over if it is auto-disabled. If you want to override this behavior, you can use the autoenable option to clear the AutoDisabled flag for a service group on a system. For example, "hagrp -autoenable foo -sys bar" will clear the AutoDisabled flag for the service group foo on system bar.
Caution: To bring a group online manually after VCS has autodisabled it, make sure that the group is not fully or partially active on any system that has the AutoDisabled attribute set to 1 by VCS. Specifically, verify that all resources that may be corrupted by being active on multiple systems are brought down on the designated systems. Then, clear the AutoDisabled attribute for each system:# hagrp -autoenable service_group -sys system_name
-enableresources group
  Enable all resources in a service group. Agents monitor the resources in the group.
-disableresources group
  Disable all resources in a service group. Agents do not monitor the resources in the group.
-ignore group -clus cluster
  The -ignore option marks the selected remote group as offline on the specified remote cluster. The remote cluster specified should be in the \(oqfaulted\(cq state. This option is used in conjunction with the -status and -failover options of the haclus command. This ignores groups from the faulted cluster as seen from the local cluster. Ignored groups will not be listed in the output of the -status and -failover options of the haclus command.
-display [group(s)] [-attribute attribute(s)] [-sys system(s)] [-clus
  cluster | -localclus]
Display the attributes and their values for a specified group (or all groups if no group is specified). If the system is specified, display the attributes and values for the specified group on the specified system. If the cluster is specified, display the attributes and values for the specified group on the specified cluster. If the Global Cluster option is in use, the command displays the specified information for all systems on all clusters. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. For a system whose name is not unique among clusters, specify the name of its cluster. Only the values of attributes of global groups can be displayed across clusters.
-state [group(s)] [-sys system(s)] [-clus cluster | -localclus]
  Display the current state of the specified groups on the specified system (or all systems if no system is specified) in the specified cluster. If the Global Cluster option is in use, the command lists the state of the specified system. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. For a system whose name is not unique among clusters, specify the name of its cluster. Only the states of global groups can be displayed across clusters.
-list [conditional(s)] [-clus cluster | -localclus]
  Displays a list of groups whose values match given conditional statement(s). Conditional statements can take three forms: Attribute=Value, Attribute!=Value, Attribute=~Value. Multiple conditional statements imply AND logic. If no conditional statement is specified, all groups in the cluster are listed. If the Global Cluster option is in use, the command lists the specified global groups on all clusters. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. Only global groups can be displayed across clusters.
-value group attr system [-clus cluster | -localclus]
  The -value option provides the value of a single group attribute. For example, "hagrp -value groupX State sysb -clus C2" displays the value of the \(oqState\(cq attribute for the group "groupX" on system "sysb" in the remote cluster C2. If no cluster is specified, the local cluster is considered. The system name must be specified for local attribute values but not for global attribute values. The -value option is used instead of the -display option when one specific attribute value is needed rather than a table of many attribute values. If the Global Cluster option is in use, the command displays the values for the attribute on all clusters. Use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued. For a system whose name is not unique among clusters, specify the name of its cluster. Only a global group’s attribute value can be displayed across clusters.
-dep [group(s)]
 

Display dependencies between groups.

-resources group [-clus cluster | -localclus]
  List resources for a service group. For global clusters, use the -clus or -localclus option to indicate a specific cluster. The -localclus option specifies the cluster in which the command is issued.
-wait group attribute value [-clus cluster] [-sys system] [-time
  timeout]
The -wait option is for use in scripts to direct the hagrp command to block until the value of the attribute is changed as specified, or until the timeout expires. If a timeout is not specified, the command line interface blocks indefinitely. timeout is an integer specifying seconds. See EXAMPLES. The -wait option can be used only with changes to scalar attributes. The -sys option can be applied only when the scope of the attribute is local. The -clus option applies only when applying the -wait option in a Global Cluster environment. The scalar group level attributes on the remote cluster are limited to those capable of being displayed using the hagrp -display -clus remote_cluster command.
-help [-modify | -link | -list]
  Display information about using hagrp. The -modify option provides modify specific help; the -link option provides link specific help; the -list option provides list specific help.
-modify modify_options
  The -modify option lets you modify a service group’s attributes. You must make the configuration read-write before modifying any attributes. Some attributes, such as ProbesPending, are internal to VCS and cannot be modified. You can modify any attribute that can be configured in main.cf. You may modify a scalar attribute’s existing value. You may not use -modify to change values already defined for a vector, a keylist, or an association attribute. For vector, keylist, and association attributes, use one of the modify_options, which include -add, -delete, -update, or -delete -keys. Refer to the following list of permissible -modify commands. You may display the commands by using # hagrp -help -modify.
SCALAR hagrp -modify group attr value [-sys system]
VECTOR Use the following command only when the attribute has no value:

hagrp -modify group attr value ... [-sys system]

For vector attributes that have values defined, only the following operations are allowed:

hagrp -modify group attr -add value ... [-sys system]

hagrp -modify group attribute -insert index value ... [-sys system]

You can use the -insert option to add one or more values at a given index. You must enter a valid integer for index that can range from 1 to the maximum number of values.

hagrp -modify group attr -delete key ... [-sys system]

hagrp -modify group attr -delete -keys [-sys system]

KEYLIST
  Use the following command only when the attribute has no value:

hagrp -modify group attr key ... [-sys system]

For keylist attributes that have values defined, only the following operations are allowed.

hagrp -modify group attr -add key ... [-sys system]

hagrp -modify group attribute -insert index key ... [-sys system]

You can use the -insert option to add one or more values at a given index. You must enter a valid integer for index that can range from 1 to the maximum number of values that exist for the key.

hagrp -modify group attr -delete key ... [-sys system]

hagrp -modify group attr -delete -keys [-sys system]

ASSOCIATION
  Use the following command only when the attribute has no value:

hagrp -modify group attr {key value} ... [-sys system]

For association attributes that have values defined, only the following operations are allowed.

\fBhagrp -modify group attr -add {key value} ... [-sys system]

hagrp -modify group attr -update {key value} ... [-sys system]

hagrp -modify group attr -delete key ... [-sys system]

hagrp -modify group attr -delete -keys [-sys system]

-migrate service_group -to system
  Migrate a service group from a system to another system.

The migrate command is only applicable for service groups that contain resources that support live migration.

EXAMPLES

To online group db_grp on system mars01, enter:

     # hagrp -online db_grp -sys mars01

From a script, to use the -wait option to direct the hagrp command to block until the service group db_grp comes ONLINE on system mars01, enter:

     # hagrp -wait db_grp State ONLINE -sys mars01

Parent group db_grp has online local firm dependency on child group vol_grp, which has online local firm dependency on child group dk_grp configured on systems mars01, mars02, and mars03. To bring parent group db_grp and all of its required child groups online on system mars01, enter:

     # hagrp -online -propagate db_grp -sys mars01

Parent group db_grp has online local firm dependency on group vol_grp, which has online local firm dependency on group dk_grp configured on systems mars01, mars02, and mars03. To bring child group dk_grp and all of its dependent parent groups offline, enter:

     # hagrp -offline -propagate dk_grp -any

A sample command output of # hagrp -forecast group\fB -verbose:

The following output is for a service group without virtual machine resources.

Target Node         node 1
Online Node(s)      
FailOverPolicy      BiggestAvailable
Total MeterWeight    16.00
                    CPU(Units)  Mem(Units) Swap(Units)    
MeterWeight           10           5           1              
                    CPU(CPU)    Mem(MB)    Swap(MB)       
Load                 2.00        1200.00   1200.00        
Forecasted Free/Total Capacity(%free)       
Systems CPU(CPU)        Mem(MB)               ....
*node 1 3.96/4.00(99)%  1276.00/2896.00(44)%  
node 2  4.00/4.00(100)% 1331.00/2896.00(45)%  
node 3  3.98/4.00(99)%  1284.00/2896.00(44)% 
          Swap (MB)        SCPU (CPU)    SMEM (MB)                                                                                                                                                                                                      
    2553.00/2560.00(99)%                   -                                                                     -
    1529.00/1536.00(99)%                   -                                                                            -                                                                                                               
    1529.00/1536.00(99)%                   -                                                                            -
                                                                        Loads Met        SysWeight/Total MeterWeight
                                                                         Yes                 15.69/16.00
                                                                         Yes                 15.60/16.00  
                                                                         Yes                                                                                                                                    15.37/16.00                                                                                                                                                   

A sample command output of # hagrp -forecast VirtualMachineGroup -policy BiggestAvailable -verbose

The following output is for a service group with virtual machine resources.

Target Node            Node1
Online Node(s)         Node2
FailOverPolicy         BiggestAvailable
Total MeterWeight      20.00
                        CPU(Units) Mem(Units) Swap(Units)  SCPU(Units)     SMem(Units)
MeterWeight             -           -          -                        10              10
                       CPU(CPU)   Mem(MB)   Swap(MB)      SCPU(CPU)       SMem(MB)
Load                    -           -          -            12.00           4096.00
                        Forecasted Free/Total Capacity(%free)
Systems                 CPU(CPU)                Mem(MB)            Swap(MB)                                                        
*Node1                    -                       -                     -                               
 Node3                    -                       -                     -                                    
                        SCPU(CPU)               SMem(MB)
                        47.00/128.00(36%)       17600.00/65526.00(26%)
                        12.00/128.00(9%)        9504.00/65526.00(14%)
                        Loads Met       SysWeight/Total MeterWeight
                        Yes             20.00/20.00
                        Yes             7.95/20.00

A sample command output of # hagrp -forecast group :

Target Node   node 1
Online Node(s)            -
FailOverPolicy            BiggestAvailable

NOTES

The VCS engine may reject some hagrp commands. For example, VCS does not allow you to online a failover service group on a system if the group is online elsewhere in the cluster, or if the group is faulted on that system.

When using the command to specify or modify an attribute’s value that begins with a dash ("-"), precede the value with a percent sign ("%"). For example, specify -y as %-y. Likewise, precede a value that starts with a percent sign with another percent sign.

SEE ALSO

hares(1M), haclus(1M), haconf(1M)

COPYRIGHTS

Copyright (c) 2016 Veritas.

All rights reserved.


VCS 7.2 HAGRP(1M)