Update design for garbage collection and version vector #1019

JOOHOJANG · 2024-09-26T10:55:45Z

What this PR does / why we need it:
Since we're gonna use version vector to run GC after this.
I just updated design document for garbage collection and version vector.

Which issue(s) this PR fixes:

Fixes #

Special notes for your reviewer:

Does this PR introduce a user-facing change?:

Additional documentation:

Checklist:

Added relevant tests or not required
Didn't break anything

coderabbitai · 2024-09-26T10:55:51Z

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media?

❤️ Share

🪧 Tips

Chat

There are 3 ways to chat with CodeRabbit:

Review comments: Directly reply to a review comment made by CodeRabbit. Example:
- I pushed a fix in commit <commit_id>, please review it.
- Generate unit testing code for this file.
- Open a follow-up GitHub issue for this discussion.
Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
- @coderabbitai generate unit testing code for this file.
- @coderabbitai modularize this function.
PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
- @coderabbitai gather interesting stats about this repository and render them as a table. Additionally, render a pie chart showing the language distribution in the codebase.
- @coderabbitai read src/utils.ts and generate unit testing code.
- @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.
- @coderabbitai help me debug CodeRabbit configuration file.

Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments.

CodeRabbit Commands (Invoked using PR comments)

@coderabbitai pause to pause the reviews on a PR.
@coderabbitai resume to resume the paused reviews.
@coderabbitai review to trigger an incremental review. This is useful when automatic reviews are disabled for the repository.
@coderabbitai full review to do a full review from scratch and review all the files again.
@coderabbitai summary to regenerate the summary of the PR.
@coderabbitai resolve resolve all the CodeRabbit review comments.
@coderabbitai configuration to show the current CodeRabbit configuration for the repository.
@coderabbitai help to get help.

Other keywords and placeholders

Add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.
Add @coderabbitai summary to generate the high-level summary at a specific location in the PR description.
Add @coderabbitai anywhere in the PR title to generate the title automatically.

CodeRabbit Configuration File (`.coderabbit.yaml`)

You can programmatically configure CodeRabbit by adding a .coderabbit.yaml file to the root of your repository.
Please see the configuration documentation for more information.
If your editor has YAML language server enabled, you can add the path at the top of this file to enable auto-completion and validation: # yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json

Documentation and Community

Visit our Documentation for detailed information on how to use CodeRabbit.
Join our Discord Community to get help, request features, and share feedback.
Follow us on X/Twitter for updates and announcements.

emplam27

What a great document! it's good to understand about VV and GC👍👍👍

emplam27 · 2024-10-02T15:55:47Z

design/version-vector.md

+
+A Version Vector is a mechanism used in distributed systems to track the causal relationships between different replicas or nodes. Each node maintains a vector, where each element corresponds to a version counter for every other node in the system, including itself. Whenever a node updates its state, it increments its own counter. When nodes exchange information, they merge their vectors by taking the maximum value for each element, ensuring that the system knows which updates happened concurrently or which one came first. Version Vectors are particularly useful for detecting conflicts and maintaining consistency in distributed databases by providing a more detailed view of the system's state than a single logical clock.
+
+![version-vector-example](media/version-vector-example.jpg)


I cannot understand about this traditional version vector image version-vector-example.jpg. From A@2 to B@3, why the B@3 value is [2, 2, 0] not [2, 1, 0]? I understood version vector From A@2 to B@3 is work like [max(2, 0), max(0, 0)+1, max(0, 0)].

Whenever a node updates its state, it increments its own counter. When nodes exchange information, they merge their vectors by taking the maximum value for each element,

This part can be confusing because the case that you've mentioned is Vector Clock.
Also, implementation can be diverged but the important part is maintaining the concept of Version Vector.
version at B@3 is always identical at any time.

For your information, here is a specific explanation. A's lamport increases at A@2, and when B gets A@2, it'll be max([2, 0, 0], [0, 0, 0]) => [2, 0, 0], then we need to compute B's new lamport which is max(2, 1).

thx! understand it!

emplam27 · 2024-10-02T15:55:55Z

design/garbage-collection.md

 ### How It Works

 Garbage collection checks that deleted nodes are no longer referenced remotely and purges them completely.

-Server records the logical timestamp of the last change pulled by the client whenever the client requests PushPull. And Server returns the smallest logical timestamp, `min_synced_seq` of all clients in response PushPull to the client. `min_synced_seq` is used to check whether deleted nodes are no longer to be referenced remotely or not.
+Server records the version vector of the last change pulled by the client whenever the client requests PushPull. And Server returns the vin version vector, `minVersionVector` of all clients in response PushPull to the client. `minVersionVector` is used to check whether deleted nodes are no longer to be referenced remotely or not.


Is 'vin' a typo for 'min'?

And Server returns the vin version vector, `minVersionVector` of all clients in response PushPull to the client.

Oh my bad. I'll fix it

emplam27 · 2024-10-02T16:14:50Z

design/garbage-collection.md

+
+After client receive this minVersionVector, it will filter its version vector to remove detached client's lamport.
+The next pushpull request will contains filtered version vector so that eventually db.version vector will store attached client's version vector only.
+![filter-version-vector](media/filter-version-vector.jpg)


I think the image is different from the code above. If the image sync with the code above, It would be easier to understand!

This code filtered c1's vv. i think this code will be filterd c3's vv

minVersionVector = {c1:3, c2:3, c3:5}.Filter([c1]) = {c1:3, c2:3}

emplam27 · 2024-10-02T16:26:01Z

design/garbage-collection.md


-Meanwhile, client `c2` inserts `"c"` after textnode `"b"`.
+Meanwhile, `client b` inserts `"c"` after textnode `"b"`.


client b's change is local change, so there are not send change 2b to server and it changes only in local. Am I understanding this correctly?

You're right. client b's change has not been sent in that moment(so it stays as local change).
Only client a's change has been sent.

How about explaining local change for unfamiliar member to yorkie?

emplam27 · 2024-10-02T16:38:42Z

design/garbage-collection.md

+=> [c1:2, c2:1, c3:4, c4:0]
+
+```
+### An example of garbage collection:
 #### State 1

 ![garbage-collection-1](media/garbage-collection-1.png)


How is the image's situation created? Is it like this below or just an example situation?

A updates and sends changes to the server, and at the same time, B does the same thing. This situation could happen.

emplam27 · 2024-10-02T16:42:27Z

design/garbage-collection.md


 #### State 4

 ![garbage-collection-4](media/garbage-collection-4.png)

-Finally, after client `c1` receives change `4` from server, purges textnode `"b"` because it is no longer referenced remotely.
+`Client a` pushpull but nothing to push or pull. `minVersionVector` is still `{a:1, b:1}`, so no GC happens.


client a is pull change 3b in this image. is it correct?

Yep you're right. Thanks

emplam27 · 2024-10-02T16:48:17Z

design/garbage-collection.md

+![garbage-collection-4](media/garbage-collection-5.png)
+
+`Client b` pushpull but nothing to push or pull. `minVersionVector` is now `{a:3, b:1}`, node "b" removedAt is `3@a`, and `minVersionVector[a] = 3 >= 3`, thus `client b` meets GC condition 


why client b pull minVersionVector{a:3, b:1}? Above image garbage-collection-4.png, client a pull minVersionVector{a:1, b:1}.

Ah my bad, I'll fix it
Thanks for your correction

emplam27

LGTM!😄

JOOHOJANG added 2 commits September 26, 2024 19:44

Update garbage collection design document

fdc7345

Fix typo

da5cd6a

JOOHOJANG requested a review from hackerwins September 26, 2024 10:55

JOOHOJANG self-assigned this Sep 26, 2024

hackerwins requested review from emplam27 and removed request for hackerwins September 27, 2024 06:57

JOOHOJANG added 2 commits September 28, 2024 16:25

Update version vector design document

b022d1b

Add contents to garbage collection design document

3084c7a

JOOHOJANG changed the title ~~Update design for garbage collection~~ Update design for garbage collection and version vector Sep 30, 2024

emplam27 reviewed Oct 2, 2024

View reviewed changes

Updated explanation based on review feedback

0086fc4

JOOHOJANG requested a review from emplam27 October 3, 2024 13:15

emplam27 approved these changes Oct 4, 2024

View reviewed changes

Updated explanation based on review feedback

7f29919

JOOHOJANG merged commit 30eaa9f into hybrid-clock Oct 4, 2024
1 check passed

JOOHOJANG deleted the update-design-for-garbage-collection branch October 4, 2024 02:33

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update design for garbage collection and version vector #1019

Update design for garbage collection and version vector #1019

JOOHOJANG commented Sep 26, 2024 •

edited

Loading

coderabbitai bot commented Sep 26, 2024 •

edited

Loading

Review skipped

Chat

CodeRabbit Commands (Invoked using PR comments)

Other keywords and placeholders

CodeRabbit Configuration File (`.coderabbit.yaml`)

Documentation and Community

emplam27 left a comment

emplam27 Oct 2, 2024

JOOHOJANG Oct 3, 2024

emplam27 Oct 4, 2024

emplam27 Oct 2, 2024

JOOHOJANG Oct 3, 2024

emplam27 Oct 2, 2024

emplam27 Oct 4, 2024

emplam27 Oct 2, 2024

JOOHOJANG Oct 3, 2024

emplam27 Oct 4, 2024

emplam27 Oct 2, 2024

JOOHOJANG Oct 3, 2024

emplam27 Oct 2, 2024

JOOHOJANG Oct 3, 2024

emplam27 Oct 2, 2024

JOOHOJANG Oct 3, 2024

emplam27 left a comment


		A Version Vector is a mechanism used in distributed systems to track the causal relationships between different replicas or nodes. Each node maintains a vector, where each element corresponds to a version counter for every other node in the system, including itself. Whenever a node updates its state, it increments its own counter. When nodes exchange information, they merge their vectors by taking the maximum value for each element, ensuring that the system knows which updates happened concurrently or which one came first. Version Vectors are particularly useful for detecting conflicts and maintaining consistency in distributed databases by providing a more detailed view of the system's state than a single logical clock.

		![version-vector-example](media/version-vector-example.jpg)


		Meanwhile, client `c2` inserts `"c"` after textnode `"b"`.
		Meanwhile, `client b` inserts `"c"` after textnode `"b"`.

		![garbage-collection-4](media/garbage-collection-5.png)

		`Client b` pushpull but nothing to push or pull. `minVersionVector` is now `{a:3, b:1}`, node "b" removedAt is `3@a`, and `minVersionVector[a] = 3 >= 3`, thus `client b` meets GC condition

Update design for garbage collection and version vector #1019

Update design for garbage collection and version vector #1019

Conversation

JOOHOJANG commented Sep 26, 2024 • edited Loading

coderabbitai bot commented Sep 26, 2024 • edited Loading

Review skipped

Chat

CodeRabbit Commands (Invoked using PR comments)

Other keywords and placeholders

CodeRabbit Configuration File (.coderabbit.yaml)

Documentation and Community

emplam27 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

emplam27 left a comment

Choose a reason for hiding this comment

JOOHOJANG commented Sep 26, 2024 •

edited

Loading

coderabbitai bot commented Sep 26, 2024 •

edited

Loading

CodeRabbit Configuration File (`.coderabbit.yaml`)