In order for failover to proceed, the Recovery Logs for the previous active and standby sequencers must be synchronized. The Recovery Log synchronization states are displayed by the SEQUENCER STATUS command. If the standby sequencer is in the Unsynchronized state, it does not become active. This protects against possible data loss if some write requests were not recorded in the Recovery Log for the standby sequencer. This protection is bypassed with the SEQUENCER TAKEOVER BY FORCE command, in which case the administrator must manually ensure data consistencies across both managed systems.
If the standby sequencer process is not running, its Recovery Log may fall behind. Therefore, when the standby sequencer process starts up in the Unsynchronized state, it attempts to synchronize its Recovery Log with the active sequencer starting from the last in sync checkpoint. The maximum amount of data that can transfer from the active to the standby sequencer is limited by the parameter RecoveryLogSyncBackfill.