Failover for VCS Global Cluster

Failover for VCS Global Cluster

 

Setup:

Veritas Storage foundation HA for Windows v6.1

Single node Local cluster

Single node Remote cluster

Data replicated between clusters via VVR

 

A Global cluster will only failover to the remote cluster if the entire local cluster dies

However if a Service Group fails on the local cluster it will not initiate a failover to the remote cluster.

Is there a way around this since we have single node clusters? Basically I need the cluster to automatically failover to the remote cluster if a service group fails on the local cluster.

1 Solution

Accepted Solutions
Highlighted
Accepted Solution!

Hi Marius, The default

Hi Marius,

The default ClusterFailoverPolicy is Manaul.  This means that site failover does not happen unless the GCO administrator says that it is OK. 

I've recommended changing this setting to "Connected" in the past with a few customers in similar situations as your.  By setting the policy to "Connected" failover between sites will happen if the global service group has failed on all nodes in the production site, and the GCO heartbeat is in an Alive/Running state.  This should allow a VVR migrate to be done and leave little chance for data loss.

If you set the policy to "Auto" then site failover will happen when the global servcie group fails on all nodes in the production site.  Site failover will also happen anytime that the GCO heartbeat has an issue.  This can lead to situations where the global service group is online on both sites at the same time.  This can lead to data loss.  It is not recommended to run with the ClusterFailoverPolicy set to "Auto" because of the potentical for data loss.

I hope this helps you set the ClusterFailoverPolicy to the correct settings to meet your needs.  Please let us know if you have any other questions.

Thank you,

Wally

View solution in original post

2 Replies
Highlighted

Hi, You can set

Hi,

 

You can set ClusterFailOverPolicy on the service group properties to Auto

 

Determines how a global service group behaves when a cluster faults or when a global group faults. The attribute can take the following values:

Manual - The group does not fail over to another cluster automatically.

Auto - The group fails over to another cluster automatically if it is unable to fail over within the local cluster, or if the entire cluster faults.

Connected - The group fails over automatically to another cluster only if it is unable to fail over within the local cluster.

  • Type and dimension: string-scalar

  • Default: Manual

Highlighted
Accepted Solution!

Hi Marius, The default

Hi Marius,

The default ClusterFailoverPolicy is Manaul.  This means that site failover does not happen unless the GCO administrator says that it is OK. 

I've recommended changing this setting to "Connected" in the past with a few customers in similar situations as your.  By setting the policy to "Connected" failover between sites will happen if the global service group has failed on all nodes in the production site, and the GCO heartbeat is in an Alive/Running state.  This should allow a VVR migrate to be done and leave little chance for data loss.

If you set the policy to "Auto" then site failover will happen when the global servcie group fails on all nodes in the production site.  Site failover will also happen anytime that the GCO heartbeat has an issue.  This can lead to situations where the global service group is online on both sites at the same time.  This can lead to data loss.  It is not recommended to run with the ClusterFailoverPolicy set to "Auto" because of the potentical for data loss.

I hope this helps you set the ClusterFailoverPolicy to the correct settings to meet your needs.  Please let us know if you have any other questions.

Thank you,

Wally

View solution in original post