You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Vladislav Pyatkov (Jira)" <ji...@apache.org> on 2025/12/04 19:44:00 UTC

[jira] [Created] (IGNITE-27256) Simplify server side of topology-aware RAFT client

Vladislav Pyatkov created IGNITE-27256:
------------------------------------------

             Summary: Simplify server side of topology-aware RAFT client
                 Key: IGNITE-27256
                 URL: https://issues.apache.org/jira/browse/IGNITE-27256
             Project: Ignite
          Issue Type: Improvement
          Components: replication ai3
            Reporter: Vladislav Pyatkov


h3. Motivation
The current implementation notifies the client side when a new leader appears. To avoid missing a leader, we have to refresh it for the client side on each subscription (on initialization of topology change):
{code}
PhysicalTopologyAwareRaftGroupService#refreshLeader(ClusterService, Executor, RaftGroupService, CompletableFuture<?>)
{code}
We don't need to do this if the server side sends a notification about the current leader (if the leader is in a local node).

h3. Implementation notes
The optimal way to notify about a leader after subscription is sending a term in `SubscriptionLeaderChangeRequestAcknowledge`. The term is the last when the recipient node was a leader.

h3. Definition of done
Add a notification about a current leader.
Remove the method above from the client side.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)