This section describes the storage resources and the files required to configure a volume set by using the Availability Suite feature.
Before you can define an Availability Suite volume set, you must determine the following:
The data volumes to replicate such as vol-data-paris in avsdg on cluster-paris and vol-data-newyork in avsdg on cluster-newyork.
The bitmap volume that is needed for replication, such as vol-bitmap-paris in avsdg on cluster-paris and vol-bitmap-newyork in avsdg on cluster-newyork.
One shadow volume and one bitmap shadow volume on each cluster to use for a fallback snapshot, if you choose to configure one. A fallback snapshot is a compact dependent shadow volume created on the secondary cluster immediately prior to the resynchronization of a secondary volume, from which the secondary volume can be reconstructed if resynchronization fails. One fallback snapshot can be configured for each replicated volume on each cluster.
Because the fallback snapshot is a compact dependent shadow volume, as described in the Sun StorageTek Availability Suite 4.0 Point-in-Time Copy Software Administration Guide, the shadow volume need only be large enough to contain changes to the secondary volume. For most installations a volume that is 10% the size of the secondary volume is sufficient. The bitmap shadow volume is sized according to the rules described in the Sun StorageTek Availability Suite 4.0 Point-in-Time Copy Software Administration Guide. On each cluster the shadow volume and bitmap shadow volume must be in the same device group as the replicated volume that the fallback snapshot will protect.
The logical host to use exclusively for replication of the device group avsdg, such as the logical host logicalhost-paris-1 on cluster-paris and the logical host logicalhost-newyork-1 on cluster-newyork.
All hostnames that are used for Availability Suite replication must be defined in the following locations:
The /etc/inet/hosts file on each node of each partner cluster
All naming services that are used
One device-group-volset.ini file is required for each device group that will be replicated. The volset file is located at /var/cluster/geo/avs/device-group-volset.ini on all nodes of the primary and secondary clusters of the protection group. For example, the volset file for the device group avsdg is located at /var/cluster/geo/avs/avsdg-volset.ini.
The fields in the volume set file that are handled by the Geographic Edition software are described in the following table. The Geographic Edition software does not handle other parameters of the volume set, including size of memory queue, and number of asynchronous threads. You must adjust these parameters manually by using Availability Suite commands.
|
Details on sizing the disk queue volume can be found in the Availability Suite Remote Mirror Software Administration and Operations Guide and the sndradm (1M) man page.
The Geographic Edition software does not modify the value of the Availability Suite parameters. The software controls only the role of the volume set during switchover and takeover operations.
For more information about the format of the volume set files, refer to the Availability Suite documentation and the iiadm (1M) man page.
You can automatically enable fallback snapshots to protect your replicated secondary volumes from corruption by an incomplete resynchronization as described in Protecting Data on Replicated Volumes From Resynchronization Failure. To do so, on each cluster you will configure one /var/cluster/geo/avs/device-group-snapshot.ini file for each device group whose volumes you want to protect. The device-group-snapshot.ini files are read when the device group is added to a protection group, at the same time that the /var/cluster/geo/avs/device-group-volset.ini files of the device group are read. You can also add fallback snapshots to the volumes of a device group after the device group is added to a protection group, as described in Manually Enabling Fallback Snapshots, but automatic configuration is simpler.
A fallback snapshot for one volume in a device group is enabled by using a single line in the device-group-snapshot.ini file in the following format:
master_vol shadow_vol bitmap_shadow_vol
The volumes used by the fallback snapshot are described in Availability Suite Volume Sets. The variable master_vol is the path name of the replicated volume, shadow_vol is the path name of the compact dependent shadow volume that acts as a fallback for the secondary volume, and bitmap_shadow_vol is the path name of the bitmap volume for the compact dependent shadow volume. Full path names for each volume are required, and all three volumes must be in the same device group. For a single replicated volume it is easiest to use the same volume names on each cluster, but it is not required that you do so. For example, the shadow volume on cluster-paris might be /dev/md/avsset/rdsk/d102, while the shadow volume on cluster-newyork might be /dev/md/avsset/rdsk/d108.
The following example shows one line from the /var/cluster/geo/avs/avsset-snapshot.ini file that enables a fallback snapshot on one cluster for the secondary volume /dev/md/avsset/rdsk/d100 in the device group avsset. The device group avsset was created by using Solaris Volume Manager software, but any type of device group supported by the Geographic Edition software can be used with fallback snapshots.
/dev/md/avsset/rdsk/d100 /dev/md/avsset/rdsk/d102 /dev/md/avsset/rdsk/d103
This example line contains the following types of entries:
/dev/md/avsset/rdsk/d100 – Secondary volume
/dev/md/avsset/rdsk/d102 – Fallback snapshot volume
/dev/md/avsset/rdsk/d103 – Fallback snapshot bitmap volume