feat: init PgElection #5201

CookiePieWw · 2024-12-19T04:33:34Z

I hereby agree to the terms of the GreptimeDB CLA.

Refer to a related PR or issue link (optional)

What's changed and what's your intention?

As title. Details:

For election, candidates try acquire advisory lock regularly.
- If successful, the candidate is the leader. Leader should keep alive by update leader value's expire time and acquire lock multiple times. If failed once, step down.
- Edge case:
  1. Leader hold the lock but leader value expire:
    - Happens when: leader stuck between acquire lock and renew lease;
    - Result: either a follower or the leader will delete the value and re-init a campaign.
    - Follower state: a follower finds it expire, re-init a campaign. Does not delete key to avoid conflict.
    - Leader state 1: If stuck over IDLE_SESSION_TIMEOUT, the lock is released and the campaign restart.
      - When leader come back, it find he cannot acquire the advisory lock, just become a follower.
    - Leader state 2: If stuck less then IDLE_SESSION_TIMEOUT, it will:
      - Find the leader value deleted by followers, just step down with lock released and join a new campaign,
      - Find the leader value expired, just step down with value deleted, lock released and re-init a new campaign.
  2. Leader value is valid but cannot acquire the lock.
    - Happens when: leader stuck over IDLE_SESSION_TIMEOUT and comes back.
    - Actually has been explained above in Leader state 1. The IDLE_SESSION_TIMEOUT is longer than the META_LEASE_TIME is the key that this case happens only if case 1 happens.
- If not successful, the candidate is a follower. Follower should check the leader value's expire time and acquire lock multiple times. If the leader value expired, delete the value and re-init a campaign. If get the lock, just become the leader.
For registering candidates, just regularly register themselves with expire time.

Checklist

I have written the necessary rustdoc comments.
I have added the necessary unit tests and integration tests.
This PR requires documentation updates.

coderabbitai · 2024-12-19T04:33:41Z

Important

Review skipped

Auto reviews are disabled on this repository.

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 generate docstrings to generate docstrings for this PR. (Beta)
@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.

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.

src/meta-srv/src/bootstrap.rs

CookiePieWw · 2024-12-19T04:41:29Z

src/meta-srv/src/election/postgres.rs

+    }
+
+    async fn resign(&self) -> Result<()> {
+        todo!()


Corresponding todo! to EtcdElection.

codecov · 2024-12-19T05:01:04Z

Codecov Report

Attention: Patch coverage is 0% with 21 lines in your changes missing coverage. Please review.

Project coverage is 83.81%. Comparing base (6bf574f) to head (8e70124).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5201      +/-   ##
==========================================
- Coverage   84.07%   83.81%   -0.26%     
==========================================
  Files        1181     1181              
  Lines      219334   219343       +9     
==========================================
- Hits       184405   183848     -557     
- Misses      34929    35495     +566

src/meta-srv/src/election/postgres.rs

sunng87 · 2024-12-19T06:47:16Z

The cool part of this advisory lock is it's got unlocked when the connection is down. So I think we don't need to manually unlock it.

And in order to acquire the lock, each metasrv node should call pg_try_advisory_lock periodically. The connection that holds the lock will always get t and it also acts as heartbeat to keep connection live.

CookiePieWw · 2024-12-19T06:56:00Z

The cool part of this advisory lock is it's got unlocked when the connection is down. So I think we don't need to manually unlock it.

And in order to acquire the lock, each metasrv node should call pg_try_advisory_lock periodically. The connection that holds the lock will always get t and it also acts as heartbeat to keep connection live.

Initially I think there might be situations that a metasrv holds the lock but fails to renew the lease. Since we can think holding the lock itself indicates the metasrv is still alive, we can remove the lease/expire-time part of campaign/leader value and check the lock as heartbeat instead?

update: The problem may lie in that the lease mechanism and the advisory lock are overlapped at some point: we can both check the lease/expire-time, or check the advisory lock to check the leadership. We can choose one of them or use them all.

sunng87 · 2024-12-19T08:48:40Z

Initially I think there might be situations that a metasrv holds the lock but fails to renew the lease. Since we can think holding the lock itself indicates the metasrv is still alive, we can remove the lease/expire-time part of campaign/leader value and check the lock as heartbeat instead?

Conclusion from offline discussion:

pg_try_advisory_lock can be used as both lock and heartbeat
we need to set session variable idle_session_timeout to prevent lock leak
we need to ensure leader key updated correctly by lock owner

CookiePieWw · 2024-12-20T05:33:14Z

Split to multiple patches. See #5209 for the first part

CookiePieWw added 3 commits December 19, 2024 11:43

feat: init PgElection

bb91eb5

fix: release advisory lock

903865d

fix: handle duplicate keys

7dbfb5a

CookiePieWw requested a review from MichaelScofield as a code owner December 19, 2024 04:33

CookiePieWw removed the request for review from MichaelScofield December 19, 2024 04:33

github-actions bot added the docs-not-required This change does not impact docs. label Dec 19, 2024

CookiePieWw requested a review from fengjiachun December 19, 2024 04:34

chore: update comments

a059fa1

CookiePieWw requested a review from WenyXu December 19, 2024 04:37

CookiePieWw commented Dec 19, 2024

View reviewed changes

fix: unlock if acquired the lock

0d8b8bc

sunng87 reviewed Dec 19, 2024

View reviewed changes

src/meta-srv/src/election/postgres.rs Show resolved Hide resolved

chore: add TODO and avoid unwrap

22c628f

CookiePieWw added 4 commits December 19, 2024 20:13

refactor: check both lock and expire time, add more comments

5074686

Merge branch 'main' into pg-election

522f731

chore: fmt

cb021ef

fix: deal with multiple edge cases

8e70124

CookiePieWw closed this Dec 20, 2024

CookiePieWw mentioned this pull request Dec 20, 2024

feat: init PgElection with candidate registration #5209

Merged

5 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: init PgElection #5201

feat: init PgElection #5201

CookiePieWw commented Dec 19, 2024 •

edited

Loading

coderabbitai bot commented Dec 19, 2024 •

edited

Loading

Review skipped

Chat

CodeRabbit Commands (Invoked using PR comments)

Other keywords and placeholders

Documentation and Community

CookiePieWw Dec 19, 2024

codecov bot commented Dec 19, 2024 •

edited

Loading

sunng87 commented Dec 19, 2024

CookiePieWw commented Dec 19, 2024 •

edited

Loading

sunng87 commented Dec 19, 2024

CookiePieWw commented Dec 20, 2024

feat: init PgElection #5201

feat: init PgElection #5201

Conversation

CookiePieWw commented Dec 19, 2024 • edited Loading

Refer to a related PR or issue link (optional)

What's changed and what's your intention?

Checklist

coderabbitai bot commented Dec 19, 2024 • edited Loading

Review skipped

Chat

CodeRabbit Commands (Invoked using PR comments)

Other keywords and placeholders

Documentation and Community

CookiePieWw Dec 19, 2024

Choose a reason for hiding this comment

codecov bot commented Dec 19, 2024 • edited Loading

Codecov Report

sunng87 commented Dec 19, 2024

CookiePieWw commented Dec 19, 2024 • edited Loading

sunng87 commented Dec 19, 2024

CookiePieWw commented Dec 20, 2024

CookiePieWw commented Dec 19, 2024 •

edited

Loading

coderabbitai bot commented Dec 19, 2024 •

edited

Loading

codecov bot commented Dec 19, 2024 •

edited

Loading

CookiePieWw commented Dec 19, 2024 •

edited

Loading