[SERVER-32637] Ensure that upgrading to 3.6 when on pv0 doesn't break if you don't explicitly turn off readConcernMajority Created: 10/Jan/18  Updated: 30/Oct/23  Resolved: 29/Jan/18

Status: Closed
Project: Core Server
Component/s: Replication, Storage
Affects Version/s: None
Fix Version/s: 3.6.3

Type: Bug Priority: Major - P3
Reporter: Spencer Brody (Inactive) Assignee: Eric Milkie
Resolution: Fixed Votes: 1
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Backports
Depends
Related
is related to SERVER-35941 Don't maintain full stable optime can... Closed
Backwards Compatibility: Minor Change
Operating System: ALL
Backport Requested:
v3.6
Sprint: Storage 2018-01-15, Storage 2018-01-29
Participants:
Case:

 Description   

In 3.6, readConcernMajority support is enabled by default, but having readConcernMajority support enabled while running in replication protocolVersion 0 causes problems. We should ensure that the upgrade to 3.6 works as expected even if using pv0.



 Comments   
Comment by Eric Milkie [ 29/Jan/18 ]

Note that no changes were made in the master branch, as we anticipate removing pv0 support prior to 3.7 final release.

Comment by Githook User [ 29/Jan/18 ]

Author:

{'email': 'milkie@10gen.com', 'name': 'Eric Milkie', 'username': 'milkie'}

Message: SERVER-32637 update oldest_timestamp on PV0 secondaries
Branch: v3.6
https://github.com/mongodb/mongo/commit/6e066a51683db2c367f11435e1db3f38c9c2ce50

Comment by Spencer Brody (Inactive) [ 12/Jan/18 ]

davisford@gmail.com, upgrading to protocolVersion 1 should resolve this issue, and is the recommended workaround.

This ticket is for work that will be backported to 3.6 to prevent newer versions of 3.6 from exhibiting this bad behavior on pv0, by bringing the pv0 behavior into alignment with what it was on 3.4. Note that pv0 remains deprecated in 3.6, and is planned to be removed from the next major release of MongoDB.

Comment by Davis Ford [ 11/Jan/18 ]

Question: if I upgraded from 3.4.1 => 3.6.1 and had replication protocolVersion 0, saw the startup warnings, and so I updated to protocolVersion 1 – is this still a potential issue for me on our replica set?

Comment by Spencer Brody (Inactive) [ 10/Jan/18 ]

Also make sure to test a brand new 3.6 replica set started with pv0 from the beginning.

Comment by Spencer Brody (Inactive) [ 10/Jan/18 ]

First step is to try out a basic upgrade when on pv0 and see if we repro the issue.

Generated at Thu Feb 08 04:30:49 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.