Indexing

The Indexing containers contain indexing data. There are two key pieces of information stored in the Index pod:

  • Primary key to Slot Map ID.

  • Unique secondary key to primary key mapping.

The index entries are created (an example use case) at the time of session connect and removed at the time of disconnect. Using the in-memory KV store, Indexing stores the keys in memory and provides a multi-primary write. The indexing write log (set and delete operations) are also written to Kafka pods for remote site replication. One of the Index pods writes the logs in Kafka, which are then received at the remote site. The corresponding operations are also performed at the remote site.

Note

A maximum of two vCPUs is required for deploying the Index pod in production environment.

In the event of pod failover and recovery, the Index pod recovers from:

  • Local Replica Member - The Index directly reads from the gRPC stream in bulk directly to recover data.

  • Remote Replica Member - If there is no local replica available for synchronization, the Index reads the data from the remote site instances for the same map.

    The following figures depict the Index recovery process from local and remote peers:

    Index Recovery from Local Peer
    Index Recovery from Remote Peer