High Availability Configuration

The POD will have two servers hosting the voltha, onos and NEM applications.  One server will be the active and the second will be the standby.  To maintain the system availability, if the active fails, then the standby takes over.  Here are required capabilities:

Important Data

1. All data made available to DevOps users on the active server must be available from the standby server after a failover operation. This data includes:

• historical and standing alarms
• historical and new syslogs
• recent configuration backups
• software release repositories

Status Tools

1. NEM must provide tools that give the user status of the failover operation. If there are specific programs and or processes that must be brought up, then these must be listed showing their completion status.

Failover Speed

1. The failover operation should take no more than 15 minutes to complete, at which time all features and data available on the original active server are available to users on the new active server

Backups

1. It suffices that each of the two servers in the active-standby pair back up each others’ data. If DevOps wishes to make copies and send them offline that must be supported as well.

Use the Active/Standby Feature for Software Upgrades

Since we have two identical servers that can each support the POD, we should be able to perform a software upgrade on one server while the other server supports continuing operation of the POD.

1. It must be possible for NEM to disable the automatic failover mechanism so we can upgrade the software on one side while the other side is running the POD.

1. When one side is complete, NEM should provide the interface for the user to swing the active applications over the newly-upgraded side (and back again if there are problems).

1. Then upgrade the remaining side.