My Oracle Support Banner

Solaris Cluster Overview of Failover Zones/Container - Description of Zone States (Doc ID 2070260.1)

Last updated on MAY 07, 2024

Applies to:

Solaris Cluster - Version 4.0 and later
Oracle Solaris on SPARC (64-bit)
Oracle Solaris on SPARC (64-bit)

Purpose

Describe the zone states of failover zones in Solaris Cluster 4.x.

 

Scope

Describe the Zone states when used in Solaris Cluster environment.  It concentrates on Solaris Cluster 4.x  resource type of  ORCL.ha-zone_sczbt

 

In a typical ha-zone configuration, there is also a network and storage resource within the same resource group. In order to manipulate zone's state, both network and storage resource must be online, while ha-zone resource must be disabled. 

start:

/opt/SUNWsczone/bin/control_ha-zone sczbt start -R %RS_NAME -G %RG_NAME -T %RT_NAME

Solaris Cluster expects initial state of the ha-zone to be in "configured" state on all nodes in the cluster.
Below is an  overview of commands invoked during start method:

zoneadm -z <ZONE> attach                  ; this will take the zone into "installed" state.
zoneadm -z <ZONE> boot                    ; this will take the zone into "running" state.

Additionally, start method Waits for SMF service Milestone to be reached before declaring the service online.
Milestone is defined in parameter Milestone. By default is set to svc:/milestone/multi-user-server

 

 When a zone state is not in configured state, an error like below is logged.

... node1 SC[SUNWsczone.start_sczbt]:zone-rg:zoneboot-rs: [ID 812858 daemon.error] Function: start_sczbt - Zone Z1 is in state unavailable on node node2 - unable to attach zone on this node.

 

Running zoneadm list -cv will give the state of the zone on particular node.   See System Administration Guide: Oracle Solaris Containers-Resource Management and Oracle Solaris Zones  for more information. The common sub-commands of zoneadm that would used to correct the zone state is detach, attach and boot.  In particular cases,  a zone could be in incomplete state.  This requires run zoneadm's uninstall command which could destroy data. The best chance of avoiding data destruction is to ensure the zonepath is not mounted on the host when this uninstall command is run. 

 

 stop:

(/opt/SUNWsczone/bin/control_ha-zone sczbt stop -R %RS_NAME -G %RG_NAME -T %RT_NAME

zlogin <ZONE> /usr/sbin/shutdown      ; brings the zone to installed state
zoneadm -z <ZONE> detach                  ; brings the zone to configured state

 

probe:

 /opt/SUNWsczone/bin/control_ha-zone sczbt probe -R %RS_NAME -G %RG_NAME -T %RT_NAME

zlogin <ZONE> /bin/svcs -H -s -o STATE Milestone     ; run the command svcs to check the zone is correct Milestore.
                                                                              ; by default Milestone is set to svc:/milestone/multi-user-server

 When probe times out, it is usually related to performance or bugs.  Therefore, it would be a good step to ensure that they have the latest SRU.

 

 Kernel zone:

As of Solaris 11.2 and Solaris Cluster 4.2, Kernel zone may be used as Failover container.  The installation steps are discussed in article How to Configure a Failover Oracle Solaris Kernel Zone.  Kernel zone provides a more isolated environment as it has its own packaging.  It also allows for warm migration of the zone between Cluster nodes. Since the memory of the zone is also migrated, the migration time is dependent on memory being used in zone.

 

Details

To view full details, sign in with your My Oracle Support account.

Don't have a My Oracle Support account? Click to get started!

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.