[DOCS-13482] Investigate changes in SERVER-46288: Reconfig in 4.2 style with the current config on FCV downgrade Created: 02/Mar/20 Updated: 13/Nov/23 Due: 20/Mar/20 Resolved: 19/Mar/20 |
|
| Status: | Closed |
| Project: | Documentation |
| Component/s: | manual |
| Affects Version/s: | None |
| Fix Version/s: | 4.3.4, Server_Docs_20231030, Server_Docs_20231106, Server_Docs_20231105, Server_Docs_20231113 |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Backlog - Core Eng Program Management Team | Assignee: | Kay Kim (Inactive) |
| Resolution: | Works as Designed | Votes: | 0 |
| Labels: | docs-administration | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||||||||||||||
| Participants: | |||||||||||||||||||||
| Days since reply: | 3 years, 46 weeks, 6 days ago | ||||||||||||||||||||
| Epic Link: | DOCS: 4.4 Server Release Work | ||||||||||||||||||||
| Description |
DescriptionDownstream Change Summary FCV downgrade from 4.4 to 4.2 does a reconfig automatically and waits for the new config to replicate on all nodes in the replica set, including non-voting nodes and arbiters. Description of Linked Ticket Safe reconfig introduces a new field on the config which will make 4.2 binary to fail to start up. Now that we allow 4.4 binary to use and store a config with the new field. We cannot just remove the extra field on downgrade on each node as in Instead, we'll do a reconfig with a higher config version but no term if the last config has a term field with the same configuration and wait for it to replicate to all nodes before downgrading FCV. Also remove the FCV check in ReplSetHeartbeatArgsV1::addToBSON, we don't need to check FCV to decide whether to include "configTerm" in heartbeats. Scope of changesImpact to Other DocsMVP (Work and Date)Resources (Scope or Design Docs, Invision, etc.) |
| Comments |
| Comment by Kay Kim (Inactive) [ 19/Mar/20 ] |
|
Thanks much. I think we're all set w/o needing extra documentation since on: where we already
|
| Comment by Siyuan Zhou [ 19/Mar/20 ] |
|
This is separate from the proposal of arbiter work which we decided not to do. We run a reconfig automatically on FCV downgrade and wait for it to replicate to all nodes. That's a stronger requirement than waiting for FCV to replicate to a majority of nodes, but we already expect all nodes to be alive on FCV downgrade, so this behavior may not noticeable by users. I'd trust Doc team's judgement on whether this is worth documenting. |