Skip to content

Conversation

nick1udwig
Copy link
Member

Problem

providerfren.os still freezes

Solution

To solve the cause of the freeze:

  1. dont unsubscribe, since subs are de-duped by alloy; if you unsub one mint, e.g., you unsub all
  2. cache eth requests
  3. retry eth requests once on rpc error

It is still unsettling that not handling unsubs properly leads to a freeze, and that needs to be run down, but these changes seem to solve the cause.

Testing

See if providerfren.os is still alive (it is running this runtime version)

Docs Update

None

Notes

None

nick1udwig and others added 2 commits August 30, 2024 00:45
1. dont unsubscribe, since subs are de-duped by
   alloy; if you unsub one mint, e.g., you unsub all
2. cache eth requests
3. retry eth requests once on rpc error
@nick1udwig nick1udwig requested a review from barraguda August 30, 2024 07:48
Copy link
Contributor

@barraguda barraguda left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nice!

I think something we can do in the near future to make the caching more robust (and for potential further deduplication of eth requests) is to specify our get_logs requests more thoroughly.

Saving state in kns & app_store probably too, but mainly getting a block number at boot, then specifying from_block -> to_block in the filter.

Base automatically changed from hf/increase-event-loop-queue-size to develop August 30, 2024 23:33
@nick1udwig nick1udwig merged commit e1923d9 into develop Aug 30, 2024
@nick1udwig nick1udwig deleted the hf/fix-eth-provider branch August 30, 2024 23:36
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants