-
Notifications
You must be signed in to change notification settings - Fork 4.8k
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
config: reinstate #8478 (unification of delta and SotW xDS), reverted by #8939 #8974
Conversation
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>
@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 |
Sure, waiting for the CI to pass. |
Signed-off-by: Fred Douglas <fredlas@google.com>
…pcmux Signed-off-by: Fred Douglas <fredlas@google.com>
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>
@kyessenov looks good for a quick smoke test. Thank you! cc @wgallagher /wait-any |
I ran go-control-plane CI test, and it passes fine now. Thanks. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks!
Awesome, thanks. I'm relieved it got back in so quickly! |
@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 |
@rgs1 if it's causing issues we will revert again, though a bit of debugging would be nice. Can you get any more info? |
@mattklein123 yeah, any ideas on what info would be useful? |
Logs? Which xDS? Etc.? If you can reliably reproduce any/all info would be helpful. |
It might have to wait until later/Mon; do if it already caused issues a
revert might be appreciated so that we can fetch some other stuff from
master. Thanks!
…On Fri, Nov 15, 2019, 6:38 PM Matt Klein ***@***.***> wrote:
Logs? Which xDS? Etc.? If you can reliably reproduce any/all info would be
helpful.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#8974?email_source=notifications&email_token=AAKJSA6YTMRS3W4ICZKCKOLQT4JDRA5CNFSM4JL3GBSKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEGZLFQ#issuecomment-554538390>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAKJSA6T7FXSZRTTW6LBEU3QT4JDRANCNFSM4JL3GBSA>
.
|
Are you positive that if you revert the change the problem goes away? |
Positive -- tho honestly it's probably fine to wait until EOD Monday to
revert if we can't figure it out.
…On Fri, Nov 15, 2019, 6:41 PM Matt Klein ***@***.***> wrote:
Are you positive that if you revert the change the problem goes away?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#8974?email_source=notifications&email_token=AAKJSAZL2ZZR7FD2VKQQ7R3QT4JPBA5CNFSM4JL3GBSKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEGZSFY#issuecomment-554539287>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAKJSA5GL2GL2UCW7IOJHITQT4JPBANCNFSM4JL3GBSA>
.
|
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. |
…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>
I opened #9044 and we can decide on Monday whether to merge or not. |
…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>
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