Page MenuHomePhabricator

[native] Deprecate socket crash loop recovery specialized to policy acknowledgment
ClosedPublic

Authored by ashoat on Mon, Jun 10, 9:07 PM.
Tags
None
Referenced Files
Unknown Object (File)
Sun, Jun 30, 12:09 AM
Unknown Object (File)
Wed, Jun 26, 5:09 AM
Unknown Object (File)
Tue, Jun 25, 3:03 PM
Unknown Object (File)
Sat, Jun 22, 8:39 AM
Unknown Object (File)
Sat, Jun 22, 4:41 AM
Unknown Object (File)
Wed, Jun 19, 8:44 PM
Unknown Object (File)
Wed, Jun 19, 11:22 AM
Unknown Object (File)
Tue, Jun 18, 6:42 PM
Subscribers

Details

Summary

We used to have a special way to recover from a socket crash loop that results from policy acknowledgment.

Now that we have a more general way to recover from a socket crash loop, we can deprecate this old code, and rely on the more general mechanism going forward.

Test Plan

I was inspired by @kamil's test plan in D6250:

  1. Log out a test user
  2. DELETE FROM policy_acknowledgments WHERE user = '<userID>';
  3. Apply this single-line patch to force the keyserver to timeout all socket initializations: https://gist.github.com/Ashoat/c5d2d81dcb8517595f28fdff6604daa9
  4. Log in as the test user
  5. Confirm that we can heal from the simulated crash loop via fetch_pending_updates

Diff Detail

Repository
rCOMM Comm
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

This revision is now accepted and ready to land.Tue, Jun 11, 2:28 AM