Sharing a Highly Available Local File System Across Zone Clusters

You can use the SUNW.HAStoragePlus resource type to share a highly available file system directory managed by a global cluster resource to a zone cluster. This method consolidates the storage and shares a highly available local file system with different applications running on different zone clusters. For information on adding a file system to a zone cluster, see Adding File Systems to a Zone Cluster in Oracle Solaris Cluster Software Installation Guide.

This section explains the requirements and procedures for sharing a highly available local file system directory across zone clusters.

Configuration Requirements for Sharing a Highly Available Local File System Directory to a Zone Cluster

The directory of a highly available local file system managed by a global cluster resource can be shared to a zone cluster. To share a highly available local file system directory, the configuration must meet the following requirements:

• Create an HAStoragePlus resource in a failover resource group in a global cluster with the file system where the directory to be shared belongs.

• The directory of the highly available local file system that you want to share must be configured to a zone cluster as an lofs file system.

• Create an HAStoragePlus resource in a failover resource group in a zone cluster with the lofs file system.

• The zone cluster resource must have an offline restart dependency on the global cluster resource.

• The zone cluster resource's resource group must have a strong positive affinity or strong positive affinity with failover delegation on the global cluster resource's resource group.

How to Set Up the HAStorage Plus Resource Type to Share a Failover File System Directory to a Zone Cluster

The following procedure explains how to set up the HAStoragePlus resource type to share a failover file system (for example, UFS, QFS, or a ZFS pool directory to a zone cluster called zoneclustername.

1. On any node in the global cluster, become superuser or assume a role that provides solaris.cluster.modify RBAC authorization. Perform the steps from a node in the global cluster, because the dependencies and affinities from a zone cluster to a global cluster can only be set by an authorized global cluster node administrator.
2. Create a failover resource group in the global cluster.

   # clresourcegroup create gc-hasp-resource-group

3. Register the HAStoragePlus resource type in the global cluster.

   # clresourcetype register SUNW.HAStoragePlus

4. Create an HAStoragePlus resource in a failover resource group of the global cluster with the failover file system that contains the directory that you want to share to a zone cluster.

   # clresource create -g gc-hasp-resource-group -t HAStoragePlus \
   -p FilesystemMountPoints=mount-point \
   -p Zpools=pool gc-hasp-resource

5. Bring the global cluster failover resource group online in a managed state.

   # clresourcegroup online -M gc-hasp-resource-group

6. Configure the directory of the failover file system that is being shared to the zone cluster as an lofs file system.

   # clzonecluster configure zoneclustername
   clzc:zoneclustername> add fs
   clzc:zoneclustername:fs> set dir = shared-dir-mount-point-in-zc
   clzc:zoneclustername:fs> set special = shared-directory
   clzc:zoneclustername:fs> set type = lofs
   clzc:zoneclustername:fs> end
   clzc:zoneclustername> exit
   #

7. Create a failover resource group in the zone cluster that has a strong positive affinity or strong positive affinity with failover delegation on the failover resource group of the global cluster.

   # clresourcegroup create -Z zoneclustername \
   -p RG_affinities=++global:gc-hasp-resource-group \
   zc-hasp-resource-group
   OR
   # clresourcegroup create -Z zoneclustername \
   -p RG_affinities=+++global:gc-hasp-resource-group zc-hasp-resource-group

8. Register the HAStoragePlus resource type in the zone cluster.

   # clresourcetype register -Z zoneclustername SUNW.HAStoragePlus

9. Create an HAStoragePlus resource in a failover resource group of the zone cluster. Configure the zone cluster with the lofs file system for a shared directory with a dependency on the global cluster resource that you want to share to the zone cluster.

    # clresource create -Z zoneclustername -t SUNW.HAStoragePlus -g zc-hasp-resource-group \
   -p FilesystemMountPoints=shared-dir-mount-point-in-zc \
   -p Resource_dependencies_offline_restart=global:gc-hasp-resource zc-hasp-resource

10. Bring the zone cluster failover resource group online.

    # clresourcegroup online -Z zoneclustername -M zc-hasp-resource-group

Example 2-42 Setting Up the HAStoragePlus Resource Type to Share a UFS Failover File System Directory to a Zone Cluster

The following example shows how to share the /local/fs/home directory of a UFS failover file system (/local/fs) to a zone cluster called sczone.

# clresourcegroup create gc-hasp-rg
# clresourcetype register -Z sczone SUNW.HAStoragePlus
# vi /etc/vfstab /dev/md/dg1/dsk/d0 /dev/md/dg1/rdsk/d0 /local/fs ufs 2 no logging
# clresource create -g gc-hasp-rg -t SUNW.HAStoragePlus \
-p FilesystemMountPoints=/local/fs gc-hasp-rs
# clresourcegroup online -M gc-hasp-rg

The steps above ensure that the gc-hasp-rs resource running in the global cluster manages the failover file system /local/fs.

# clzonecluster configure sczone
clzc:sczone> add fs
clzc:sczone:fs> set dir = /share/local/fs/home
clzc:sczone:fs> set special = /local/fs/home
clzc:sczone:fs> set type = lofs
clzc:sczone:fs> end
clzc:sczone> exit

The configuration above makes the failover file system's directory /local/fs/home available in the zone cluster sczone at mount point /share/local/fs/home.

# clresourcegroup create -Z sczone \
-p RG_affinities=++global:gc-hasp-rg zc-hasp-rg
# clresourcetype register -Z sczone SUNW.HAStoragePlus
# clresource create -Z sczone -t HAStoragePlus -g zc-hasp-rg \
-p FilesystemMountPoints=/share/local/fs/home \
-p Resource_dependencies_offline_restart=global:gc-hasp-rs zc-hasp-rs 
# clresourcegroup online -Z sczone -M zc-hasp-rg

The steps above create a zone cluster resource that manages the shared directory as an lofs file system. The steps in this example are applicable to VxFS and QFS file systems.

Example 2-43 Setting Up the HAStoragePlus Resource Type to Share a ZFS Pool Directory to a Zone Cluster

The following example shows how to share the ZFS pool "tank" directory /tank/home to a zone cluster called sczone.

# clresourcegroup create gc-hasp-rg
# clresourcetype register SUNW.HAStoragePlus
# clresource create -g gc-hasp-rg -t SUNW.HAStoragePlus \
-p Zpools=tank gc-hasp-rs
# clresourcegroup online -M gc-hasp-rg

The steps above ensure that the ZFS failover file system is managed by gc-hasp-rs running in the global cluster.

# clzonecluster configure sczone
clzc:sczone> add fs
clzc:sczone:fs> set dir = /share/tank/home
clzc:sczone:fs> set special = /tank/home
clzc:sczone:fs> set type = lofs
clzc:sczone:fs>end
clzc:sczone> exit
#

The configuration above makes the ZFS pool "tank" directory /tank/home available in the zone cluster sczone at mountpoint /share/tank/home.

 # clresourcegroup create -Z sczone \
-p RG_affinities=++global:gc-hasp-rg zc-hasp-rg
# clresourcetype register -Z sczone SUNW.HAStoragePlus
# clresource create -Z sczone -t HAStoragePlus -g zc-hasp-rg \
-p FilesystemMountPoints=/share/tank/home \
-p Resource_dependencies_offline_restart=global:gc-hasp-rs zc-hasp-rs
# clresourcegroup online -Z sczone -M zc-hasp-rg

The steps above create a zone cluster resource that manages the shared directory as an lofs file system.