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

config: reinstate #8478 (unification of delta and SotW xDS), reverted by #8939 #8974

Merged
merged 6 commits into from
Nov 12, 2019

Conversation

fredlas
Copy link
Contributor

@fredlas fredlas commented Nov 11, 2019

Delta and state-of-the-world gRPC xDS now share the same GrpcMux and
Subscription implementations. These are the implementations added in #7293 for
delta. These implementations are a rewrite/extensive refactoring of the old
GrpcMuxImpl (deleted by this PR) and associated logic.

Risk Level: high (rewrite/large refactor of gRPC xDS)
Testing: added sotw_subscription_state_test, added regression test for #8927

Signed-off-by: Fred Douglas <fredlas@google.com>
@repokitteh-read-only
Copy link

CC @envoyproxy/api-shepherds: Your approval is needed for changes made to api/.

🐱

Caused by: #8974 was opened by fredlas.

see: more, trace.

@fredlas
Copy link
Contributor Author

fredlas commented Nov 11, 2019

There was a minor merge conflict with #8886 to resolve. While doing that I noticed that some comments in new_grpc_mux_impl.cc wrongly referred to "DeltaDiscoveryRequest" when they should say a generic "discovery request". I can undo that if you'd prefer it to be a pure revert revert. Other than the new_grpc_mux_impl.cc cleanup and the WellKnownGrpcStatus merge, this should be identical to #8478.

…st would have caught it

Signed-off-by: Fred Douglas <fredlas@google.com>
@mattklein123
Copy link
Member

@fredlas can you check CI?

@kyessenov once we get CI passing do you mind giving this branch a quick smoke test however you tested master last time? Thank you.

/wait

@kyessenov
Copy link
Contributor

Sure, waiting for the CI to pass.

Signed-off-by: Fred Douglas <fredlas@google.com>
…pcmux

Signed-off-by: Fred Douglas <fredlas@google.com>
@fredlas
Copy link
Contributor Author

fredlas commented Nov 12, 2019

Done, updated some more GrpcStatuses. I'm also seeing that any _speed_test test is failing to build, due to a problem when trying to link with libpthread. I'm not sure if that's a newly uncovered quirk with just my own machine's environment, or a recently introduced actual breakage that hasn't yet been caught and fixed, but it's hard to imagine something in these changes causing it (and not causing it before). I guess we'll see once the tests get there.

Signed-off-by: Fred Douglas <fredlas@google.com>
…pcmux

Signed-off-by: Fred Douglas <fredlas@google.com>
@mattklein123
Copy link
Member

@kyessenov looks good for a quick smoke test. Thank you! cc @wgallagher

/wait-any

@kyessenov
Copy link
Contributor

I ran go-control-plane CI test, and it passes fine now. Thanks.

Copy link
Member

@mattklein123 mattklein123 left a comment

Choose a reason for hiding this comment

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

Thanks!

@mattklein123 mattklein123 merged commit a37522c into envoyproxy:master Nov 12, 2019
@fredlas
Copy link
Contributor Author

fredlas commented Nov 12, 2019

Awesome, thanks. I'm relieved it got back in so quickly!

@rgs1
Copy link
Member

rgs1 commented Nov 15, 2019

@fredlas @mattklein123 hmm so apparently this change causes some of our instances to get stuck on startup -- unclear if it's CDS or SDS or something in XDS but something with XDS broke which prevents some of our binary/config configurations to start... Any ideas before I start carefully reading this PR?

cc: @fishcakez

@mattklein123
Copy link
Member

@rgs1 if it's causing issues we will revert again, though a bit of debugging would be nice. Can you get any more info?

@rgs1
Copy link
Member

rgs1 commented Nov 15, 2019

@mattklein123 yeah, any ideas on what info would be useful?

@mattklein123
Copy link
Member

Logs? Which xDS? Etc.? If you can reliably reproduce any/all info would be helpful.

@rgs1
Copy link
Member

rgs1 commented Nov 15, 2019 via email

@mattklein123
Copy link
Member

Are you positive that if you revert the change the problem goes away?

@rgs1
Copy link
Member

rgs1 commented Nov 15, 2019 via email

@mattklein123
Copy link
Member

OK I think we should probably go ahead and start the revert to get master into a clean state and then we can debug in parallel, but I don't feel that strongly about it.

mattklein123 added a commit that referenced this pull request Nov 15, 2019
…e_names_ (#8918)"

This reverts commit 80aedc1.

Revert "config: rename NewGrpcMuxImpl -> GrpcMuxImpl (#8919)"
This reverts commit 6d50553.

Revert "config: reinstate #8478 (unification of delta and SotW xDS), reverted by #8939 (#8974)"
This reverts commit a37522c.

Signed-off-by: Matt Klein <mklein@lyft.com>
@mattklein123
Copy link
Member

I opened #9044 and we can decide on Monday whether to merge or not.

mattklein123 added a commit that referenced this pull request Nov 16, 2019
…e_names_ (#8918)" (#9044)

This reverts commit 80aedc1.

Revert "config: rename NewGrpcMuxImpl -> GrpcMuxImpl (#8919)"
This reverts commit 6d50553.

Revert "config: reinstate #8478 (unification of delta and SotW xDS), reverted by #8939 (#8974)"
This reverts commit a37522c.

Signed-off-by: Matt Klein <mklein@lyft.com>
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.

4 participants