[SERVER-29436] InitialSyncer doesn't need to update optimes of ReplicationCoordinator until it completes successfully Created: 02/Jun/17  Updated: 06/Dec/22

Status: Backlog
Project: Core Server
Component/s: Replication
Affects Version/s: None
Fix Version/s: None

Type: Improvement Priority: Minor - P4
Reporter: William Schultz (Inactive) Assignee: Backlog - Replication Team
Resolution: Unresolved Votes: 0
Labels: neweng
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
related to SERVER-28677 InitialSyncer::_setup_inlock() should... Closed
Assigned Teams:
Replication
Participants:

 Description   

Currently, the InitialSyncer updates the applied op time of the ReplicationCoordinator upon every successful batch application. It may be that this is unnecessary, since, during initial sync, the values of the applied and durable op times are not relevant for any other purpose. Only once initial sync completes successfully and we transition in to steady state replication do we really need to update a node's notion of applied op time.

This change can be considered optional, but it could be a small simplification of the initial sync machinery, and be a way to reduce coupling between the InitialSyncer and the ReplicationCoordinator. This change would move the initial sync process more in the direction of being viewed an isolated state machine, that handles its own state and behaviors, independent of other replication machinery.


Generated at Thu Feb 08 04:20:52 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.