MySQL Bugs: #116027: The Problems with Paxos Variant Algorithms in Group Replication

Bug #116027	The Problems with Paxos Variant Algorithms in Group Replication
Submitted:	6 Sep 2024 11:37	Modified:	12 Sep 2024 14:04
Reporter:	Bin Wang (OCA)	Email Updates:
Status:	Verified	Impact on me:	None
Category:	MySQL Server: Group Replication	Severity:	S4 (Feature request)
Version:		OS:	Any
Assigned to:		CPU Architecture:	Any

Description:
When there are no theoretical problems with the Mencius algorithm, introducing a new Multi-Paxos algorithm brings several challenges:
1.High Maintenance Cost: Maintaining and testing two sets of codebases doubles the workload for this part.
2.Regression Testing Challenges: In practice, the new algorithm has led to several regression problems, some of which are difficult to address.
3.Partial Problem-Solving: The new algorithm may only partially address the requirements. In Group Replication’s single-primary mode, it might not be universally applicable, as consistent read and write operations require all nodes to continuously communicate information

For detailed explanations, refer to:
https://advancedmysql.github.io/blogs/paxos.html

How to repeat:
With MySQL configured to use the 'after' mechanism for strong consistency writes, running SysBench read/write tests (e.g., with 100 concurrent threads) easily reproduces the algorithm's limitations.

Hi,

Nice write-up, and interesting idea to use mysql bugs database to promote your blog but you need to put here the way to reproduce the problem

Not every issue is a bug. The Paxos family of algorithms is highly complex and requires a solid theoretical foundation. What’s being discussed here is a better solution, and some aspects require a paper-level explanation to fully understand.

Of course but this is bugs database.

There are two things we can discuss here
 - bug reports
 - contributions

So you can contribute code.

But philosophical conversation about algorithms is not. We have a forum for that.

https://forums.mysql.com/

We all visit forums regularly and devs will join in on conversation they can contribute to. Also you cal use our public MySQL slack channel.

Thanks

I spent at least six months fixing this issue, and it's now running smoothly. I can confidently say that this was a major design flaw that caused numerous problems.

It's truly disappointing to put in so much effort to raise issues, only to be told they're not bugs. Group Replication has so many problems, and it's unacceptable to ignore them. I've spent four years working with Group Replication, and I know exactly what constitutes a problem.

Thank you very much for your contribution, we appreciate it!

We verified this as Enhancement Request

In order for us to continue the process of reviewing your contribution to MySQL, please send us a signed copy of the Oracle Contributor Agreement (OCA) as outlined in https://oca.opensource.oracle.com

Signing an OCA needs to be done only once and it's valid for all other Oracle governed Open Source projects as well.

Getting a signed/approved OCA on file will help us facilitate your contribution - this one, and others in the future.  

Please let me know, if you have any questions.

Thank you for your interest in MySQL.