You are here: Core engine > High availability > High availability use case scenarios > Scenario: Performing a rolling restart

Scenario: Performing a rolling restart in a highly available system

The following steps depict a typical quiesce scenario, which includes performing a rolling restart of one or many Pega 7 Platform servers.

Using a load balancer, perform the following actions:

  1. Identify the Pega 7 Platform servers to be quiesced.
  2. Disable the identified Pega 7 Platform servers in the load balancer.
  3. Quiesce the Pega 7 Platform servers using one of the following cluster management options (for details, see Cluster management):
  4. The Pega 7 Platform servers progress through the quiesce state machine. Users passivate after periods of inactivity on the quiescing Pega 7 Platform servers and activate on new Pega 7 Platform servers in the cluster as they are sent back to the load balancer during periods of inactivity.
  5. The Pega 7 Platform servers are ready for shutdown or servicing when the system run state is "Quiesce Complete" and the active user count is 0.
  6. Perform maintenance.
  7. Restart the Pega 7 Platform servers when complete.
  8. Enable the Pega 7 Platform servers in the load balancer to start taking traffic.

If a quiesced Pega 7 Platform server is added back into the load balancer rotation, users directed to that Pega 7 Platform server receive the following error:

Status Error
Message Too many identical requests in quiesce mode. Please see your system administrator.
Operator ID Unauthenticated or not available
Requestor ID No ID available
Timestamp Thu Aug 08 12:09:20 EDT 2013
Engine Version Unknown

Note: Users do not see this message once the Pega 7 Platform server run state is "Running."

To perform a rolling restart, start again at Step 1 and repeat the process until all remaining Pega 7 Platform servers have been restarted.

Related Topics Link IconRelated information