Ownership takeover mode
Use this page to enable or disable ownership takeover mode for a failed cluster in the IBM® TS7700.
Ownership takeover mode allows other clusters in the grid to obtain
ownership of the virtual volumes owned by the failed cluster. Normally,
ownership is transferred from one cluster to another through communications
between the clusters. However, a failed cluster or communication path
prevents the normal means of transferring ownership. If the cluster
itself is no longer operational, then ownership takeover can be used
to access the virtual volumes.
Note: Do not attempt ownership takeover
if the communication path between clusters has failed but no clusters
are in a failed state. Integrity of virtual volumes in the grid can
be compromised if a takeover mode is enabled for a cluster that is
not actually in a failed state.
Read-write ownership takeover | Read-only ownership takeover |
---|---|
Operational clusters in the grid can:
|
Operational clusters in the grid can:
|
A consistent copy of the virtual volume must be available on the grid or the virtual volume must exist in a scratch category. If no cluster failure occurred, the possibility exists for two sites to write data to the same virtual volume | If no cluster failure occurred, it is possible that a virtual volume accessed by another cluster in read-only takeover mode contains older data than the one on the owning cluster. This situation can occur if the virtual volume was modified on the owning cluster while the communication path between the clusters was down. |
If the Autonomic Ownership Takeover Manager
(AOTM) is configured, it performs an additional check to determine
whether the unavailable cluster is in a failed state when an ownership
takeover is initiated. If AOTM determines that the cluster is not
in a failed state, the ownership takeover attempt fails.
- Local Cluster Summary
- This section provides a high-level summary of ownership takeover
actions performed by the accessing, or local, cluster. Information
displayed includes:
- Takeover Enabled Against
- A list of the clusters in the grid against which the accessing cluster has enabled ownership
takeover. These clusters are in the failed state and are identified using this format:
"Cluster Nickname[Cluster ID]"
(#Distributed Library Sequence Number)
- Autonomic Ownership Takeover Mode Configuration
- This area displays the current AOTM configuration for the accessing
cluster. Information displayed includes:
- Remote Cluster
- The remote, failed cluster that currently owns the virtual volumes
to be accessed. Information is displayed in this format:
"Cluster Nickname[Cluster ID]"
- Takeover Mode
- The type of ownership applied to the remote cluster. Possible values are:
- Read-Only
- Read/Write
- Grace Period (minutes)
- The amount of time the accessing cluster waits for a remote cluster to respond across the grid before attempting an AOTM handshake. This value can be between 10 and 1000 minutes. The default value is 25 minutes.
- Retry Period (minutes)
- The amount of time the accessing cluster waits between attempts to determine whether the remote cluster is online. This value can be between 5 and 100 minutes. The default value is 5 minutes.
- Configure AOTM...
- Click this button to configure the AOTM settings that are displayed
in the Autonomic Ownership Takeover Mode Configuration table.
The Configure Autonomic Ownership Takeover mode page
opens. Note: An IBM service representative must configure the TSSC IP addresses for each cluster in the grid before AOTM can be enabled and configured for any cluster in the grid.If AOTM has not been configured for a remote cluster, Remote Cluster Summary values for that cluster are displayed as "
- - -
".
- Remote Cluster Summary
- This section provides a summary of ownership takeover information for all clusters in the grid
except the accessing, or local, cluster. It also contains the Select Action
drop-down menu that permits you to manually set ownership takeover mode for a failed cluster and
another cluster in the grid has not taken ownership (AOTM was not previously configured).
- Select Action
- Use this drop-down menu to manually set ownership takeover mode for a cluster you select.
Possible actions include:
- Disable Ownership Takeover
- Choose this action to manually disable ownership takeover for the cluster you select.
- Set to Read-Only Takeover Mode
- Choose this action to manually set read-only takeover mode for the cluster you select.
- Set to Read/Write Takeover Mode
- Choose this action to manually set read-write takeover mode for the cluster you select.
Note: If a remote cluster has failed and no other clusters in the grid have taken ownership of its virtual volumes, information for that cluster occurs in bold text, sorted to appear at the top of the table.Information displayed in this table includes:- Select
- Radio buttons in this column are enabled for clusters that are eligible for user-initiated manual ownership takeover. If the radio button is not enabled for a given, remote cluster, then no actions contained by the Select Action drop-down menu can be performed on that cluster.
- Remote Cluster/SN
- The remote, failed cluster that currently owns the virtual volumes to be accessed. Information
is displayed in this format:
"Cluster Nickname[Cluster ID]"
(#Distributed Library Sequence Number)
- ID
- The cluster ID of the remote, failed cluster that currently owns the virtual volumes to be accessed.
- Connection State
- The state of the connection between the accessing cluster and remote clusters in the grid.
Possible values include:
- Failed
- The accessing cluster cannot communicate with the remote cluster because the remote cluster is in the failed state.
- Service
- The remote cluster is in the service state or is offline through the service state.
- Normal
- The remote cluster is in the normal operation state and communicating with the accessing cluster.
- Takeover State
- Whether the remote cluster has been taken over by another cluster in the grid. Possible values are:
- Disabled
- If this value is displayed in bold text and occurs at the top of the table then the cluster has failed and no other clusters in the grid have taken it over.
- Enabled
- The cluster is in the Failed state and has been taken over by the accessing cluster or another cluster in the grid.
- Current Takeover Mode
- The takeover mode in force for and established by the remote, owning cluster unless manually
changed using the Select Action drop-down menu. If the connection state is
Normal, the value in this field is "
- - -
". Other possible values are:- Read-Only
- Read/Write