HA - how to immunize against domain controller crash

[jjakub]

I am using wildfly8 and I want to have high availability and failover for one node failure - physical server down.

Domain mode is convenient for management,

but it seems to be a bottleneck - if it fails slave will not start.

Is it possible to configure slave to do a local copy of domain, and if domain controller is not available to make it start with last available configuration snapshot ?

What is a good practice for a failover with no single point of failure ?

Is recommended solution based on several independant standalone servers ?

I do not need session replication.

Thx, BR

[wdfink]

See domain start option --backup and --cached-dc. These can be used to backup the configuration and start the slave as DC if needed.

[rhusar]

Otherwise, let me note that this does not represent single point of failure since an outage of the domain controller does not affect running cluster and its HA services. It's only about management.