Lack of track IO timing gpdb feature

[reshke]

There is GPDB feature for tracking IO timing for AO/AOCS relations One can refer to implementation in GPDB archive repo https://github.com/greenplum-db/gpdb-archive/commit/fc081c758cf64ddcd526318876b171d6a2fae4e7

I faced some conflict when applying my Yezzey-related patches on CBDB HEAD, so I wonder whats is actual reason of absence of this feature. Is there some major bugs, or CBDB project does not include all GP 7 functionality by design?

Are you willing to submit PR?

• [X] Yes, I am willing to submit a PR!

Code of Conduct

• [X] I agree to follow this project's Code of Conduct.

[github-actions[bot]]

Hey, @reshke welcome!🎊 Thanks for taking the time to point this out.🙌

[gfphoenix78]

Hi @reshke , thank you for your interesting on CBDB. CBDB forked GPDB at early 2022(gp7 alpha or beta). GPDB had many commits since that time. We know the value of GPDB patches and we are porting these commits to our CBDB. We need some time to complete this work. Please continuing on focus our product, you'll see this feature in CBDB later.

[reshke]

I can actually do it myself, like cherry-pick series of patches.

However, these commit depends on out-of $subj scope work being merged in GPDB main. So, many conflict with applying as-is. Conflict are not very complicated themselves. But I wonder if there is some caveats that should be taken into account, maybe some guideline for back-porting features?

[my-ship-it]

Hi @reshke, thanks for filing this issue. Currently, the back-porting strategy is that we back-port GPDB specific commits(features and bug fixes) following the time line to get less conflicts with a commit history window of 1 - 2 months. But there is an exception for some important features, for example, resource group. We already back-ported features: https://github.com/cloudberrydb/cloudberrydb/pull/432 https://github.com/cloudberrydb/cloudberrydb/pull/448

And currently, what is going on: https://github.com/cloudberrydb/cloudberrydb/pull/531

We welcome contributions from community, if you are interested, please don't hesitate to file PR following coding contribution guide

[reshke]

We welcome contributions from community, if you are interested, please don't hesitate to file PR following coding contribution guide

HI! How we can observe state of cherry-pick project, progress and work that yet to be done?

[tuhaihe]

IMO, for the cherry-pick, it would be better to track using either a Kanban view or a table view within the GitHub project.

[tuhaihe]

Follow-ups: @reshke, I will work with @my-ship-it to draft one proposal to track the cherry picks in public. When done, will let you know, and welcome to have your feedback.