Read and write requests with Ozone Manager in High Availability

Read requests from the client applications are directed to the leader Ozone Manager (OM) node. After receiving an acknowledgement to its request, the client caches the details of the leader OM node, and routes subsequent requests to this node.

If repeated requests to the designated leader OM node start failing or fail with a NonLeaderException, it could mean that the particular node is no longer the leader. In this situation, the client must identify the correct leader OM node and reroute the requests accordingly.

The following command lists the OM leader node and the follower nodes in an Ozone cluster:
ozone admin om getserviceroles -id=<ozone service id>

In the case of write requests from clients, the OM leader services the request after receiving a quorum of acknowledgements from the follower.