Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Subtask] Provide support for Qaku #2178

Open
weboko opened this issue Oct 10, 2024 · 5 comments
Open

[Subtask] Provide support for Qaku #2178

weboko opened this issue Oct 10, 2024 · 5 comments
Assignees

Comments

@weboko
Copy link
Collaborator

weboko commented Oct 10, 2024

Description

This issue is to track all relevant problems and requests that are needed to make Qaku more usable.

@vpavlin
Copy link
Member

vpavlin commented Oct 15, 2024

Thanks for creating this!

My main issues are

  • the Filter instability
  • Store nodes returning
    Store query failed with status code: 300, description: BAD_RESPONSE: archive error: DIRVER_ERROR: error in runStmt: error in 
    dbConnQueryPrepared calling waitQueryToFinish: error in query: ERROR:  out of shared memory
    HINT:  You might need to increase max_locks_per_transaction.
    
    This is obviously problem of nodes with too little RAM, but js-waku should probably disconnect from them after a few failures?
  • Concurrent store queries leading to Cannot push value onto an ended pushable

From the perspective of Qaku itself a few things that need some attention:

  • UX when a message is published - should the button stay disabled until the message is delivered back by either Filter or Store (or should we also add more frequent Store checks based on message hash?)
  • Initial loading - there is a race where Qaku never really loads if waiting for peers take too long

@danisharora099
Copy link
Collaborator

@vpavlin is there an update on the state of stability now?

Re Waku:

  • Filter is now stable(r)
  • cc @Ivansete-status for Store SHM -- is there a permanent resolution for this, or should we cater for this on js-waku level? (perhaps should do it any way assuming the network can be unstable? :D)

@danisharora099 danisharora099 self-assigned this Nov 27, 2024
@Ivansete-status
Copy link
Contributor

@danisharora099 - I think we may need to look for another approach in some PostgreSQL queries. My bet is that this issue comes due to a temporary table that a message-hash query creates. Therefore, I see a reason for changing this approach.
Other than that, the only solution is to make sure the database host configures a bigger shm_size (I use to recommend 1GB.)

@weboko
Copy link
Collaborator Author

weboko commented Dec 3, 2024

I think we still want to make Quaku great again.
Maybe we can think of some other solution here to temporarily substitute Store, @vpavlin? That way we can get some feedback on anything but store from Town hall and such.

Also, can we integrate some telemetry into Quaku so that we can technically track behavior of js-waku?

@fryorcraken
Copy link
Collaborator

Limit time range queries to 24 hrs max, like in Status. I'd expect that to be actually provided by the messaging API.
You can still do more than one query. but avoid unbounded queries.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
Status: Code Review / QA
Development

No branches or pull requests

5 participants