Choosing a Migration Strategy

SAN-based storage can be reconfigured so the zonepath is visible on the new host.

If all of the zones on one system must be moved to another system, a replication stream has the advantages of preserving snapshots and clones. Snapshots and clones are used extensively by the pkg, beadm create, and zoneadm clone commands.

There are five steps to performing a P2V or V2V migration.

1. For P2V, analyze the source host for any Oracle Solaris configuration:

   • Determine the IP type, exclusive-IP or shared-IP, of the non-global zone based on networking requirements.

   • Determine whether any additional configuration in the global zone of the target host is required.

   • Decide how application data and file systems will be migrated.

   The zonep2vchk basic analysis performed by the -b option identifies basic issues related to Oracle Solaris configuration or features used by the source global zone. The zonep2vchk static analysis using the -s option helps identify issues related to specific applications on the source global zone. The zonep2vchk runtime analysis performed by the -r inspects the currently executing applications for operations that might not function in a zone.

2. Archive the source system or zone. This archive of the Oracle Solaris instance potentially excludes data that is to be migrated separately.

   • To archive Oracle Solaris 10 global zones, flarcreate can be used.

     See How to Use flarcreate to Create the Image.

   • To archive Oracle Solaris 10 systems and non-global zones, flarcreate with the -R or -L archiver can be used to exclude certain files from the archive. Be sure to halt the zone first.

     See How to Use flarcreate to Exclude Certain Data.

   • For Oracle Solaris 11 global zones, zfs send can be used to archive the root pool.

   • For Oracle Solaris 11 non-global zones, zfs send can be used to archive the zonepath dataset of the zone.

   • For solaris10 or solaris zones that reside in a zpool on shared storage, such as a SAN, the V2V migration strategy does not require creating an archive. SAN-based storage can be reconfigured so the zonepath is visible on the new host. To reconfigure:

     • Export, then import, the zpool on the target global zone.

     • Use zoneadm install (preferred) or attach on the target system. (See step 5 in this section.)

   Also see “Zones on Shared Storage.”

3. Choose a migration strategy for additional data and file systems, such as:

   • Include the data in the archive (see step 2 in this section).

   • Archive the data separately using a preferred archive format, such as zfs send, and restore the data in the zone after migration.

   • Migrate SAN data by accessing SAN storage from the target global zone, and making the data available to the zone by using zonecfg add fs.

   • Storage in ZFS zpools can be migrated by exporting the zpool on the source host, moving the storage, and importing the zpool on the target global zone. These ZFS file systems can then be added to the target zone using zonecfg add dataset or zonecfg add fs. Note that zpools on SAN storage devices can also be migrated in this way.

4. Create a zone configuration (zonecfg) for the target zone on the target host.

   • For P2V, use the zonep2vchk command with the -c option to assist with creating the configuration.

   • For V2V, use the zonecfg -z source_zone export command on the source host. Be sure to set the brand to solaris10 when migrating Oracle Solaris 10 Containers into Oracle Solaris 10 Zones.

   Review and modify the exported zonecfg as needed, for example, to update networking resources.

5. Install or attach the zone on the target host using the archive. A new sysconfig profile can be provided, or the sysconfig utility can be run on first boot.