Roll forward #34191, which is reverted due to error `2023-10-09
22:01:18,569 FAILED: cmake/build/client_transport_test
--gtest_filter=ClientTransportTest.AddMultipleStreams
GRPC_POLL_STRATEGY=none` (Removed uses_event_engine=False,
uses_polling=False in test build).
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
Fix: https://github.com/grpc/grpc/issues/34298
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
- Added ARM64 to MacOS to be clear about the recent Apple Silicon
support.
- Removed x86 from MacOS as x86 is no longer used for Apple.
- Changed ARM to ARM64 for Linux as Foundational-cxx-support only
support ARM64 although gRPC is still working on ARM32.
This PR aims to fix the issue found in frame_fuzzer after #34191 was
merged, where frame serialization is missing the frame header info and
causes mismatch with the original frame.
This PR needs to be merged before #34657.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
Cancel streams if we have too much concurrency on a single channel to
allow that server to recover.
There seems to be a convergence in the HTTP2 community about this being
a reasonable thing to do, so I'd like to try it in some real scenarios.
If this pans out well then I'll likely drop the
`red_max_concurrent_streams` and the `rstpit` experiments in preference
to this.
I'm also considering tying in resource quota so that under high memory
pressure we just default to this path.
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Some user reported an issue that symbols from CoreFoundation are missing
when install grpc/grpc-tools inside Conda:
* https://github.com/grpc/grpc/issues/33714
* https://github.com/grpc/grpc/issues/34135
This PR adds the `-framework CoreFoundation` flag so no workarounds are
required from user.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
Since we added support for Python 3.12, our tests have been taking
longer to complete and we've been seeing an increase in timeout errors.
We already increased timeout in #34550, this PR increase timeout for
release and pull_request jobs.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
Reverts grpc/grpc#34667
The change was reverted because it failed to import to g3, after some
changes, now it's safe to reapply those changes.
Tested by importing this PR internally, it passed presubmit:
cl/573836270
Noticed that our behavior conflicts with what's mandated:
https://www.rfc-editor.org/rfc/rfc9113.html#section-5.1.2
So here's a quick fix - I like it because it's not a disconnection,
which always feels like one outage avoided.
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
This reverts commit 601aaf80b7, which
results in rolling forward #34408.
The 601aaf reversion happened because of a deadlock found in Python. The
root cause ended up being an issue with the Python wrapper and was fixed
in #34712 , so this can be rolled forward again
Reverse dependency edge, so instead of saying `logging_filter _after_
otel`, instead say `otel _before_ logging_filter` - since this doesn't
inadvertently bring otel into builds where it's unnecessary.
Required moving filter class definitions into the header - which mirrors
all other filters, so I think this is fine.
Also required removing the bespoke visibility rules on logging_filter -
which also seems relatively fine (the defaults limit to grpc usage, and
it's hard to see a firm requirement for tighter visibility that that).
Partial rollback of https://github.com/grpc/grpc/pull/34473 for RBE as
RBE needs Ubuntu 20.04 to get clang artifacts which Google publishes.
There aren't artifacts for Ubuntu 22.04 yet so there is no point of
upgrading RBE image to 22.04.
Fix: https://github.com/grpc/grpc/issues/34672
With some recent changes in core, now `grpc_ssl_credentials_create` is
guarded by `gpr_once_init`. In our current implementation, The thread
got `gpr_once_init` lock might require GIL lock during the execution of
`grpc_ssl_credentials_create`, which might cause a deadlock if another
thread is holding GIL lock and waiting for `gpr_once_init` lock.
This change adds `with nogil` to calls to native function
`grpc_ssl_credentials_create` to make sure GIL is released before
calling `grpc_ssl_credentials_create`.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
Fixes https://github.com/grpc/grpc/issues/34482.
But this is only a bandaid, there is a bigger issue with build
dependencies. AFAICT,
https://github.com/grpc/grpc/blob/master/tools/distrib/fix_build_deps.py
does not work on many targets, e.g. if the target has specified the
`nofixdeps` tag or if a field is set to a variable, e.g.
```
GRPCXX_SRCS = [
"a",
"b",
"c",
]
...
srcs = GRPCXX_SRCS,
```
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
Ditch the old priority scheme for ordering filters, instead explicitly
mark up before/after constraints.
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Introduce `grpc-native-debug` gems containing debug symbol packages that
complement the shared libraries shipped in pre-compiled binary gems.
src/ruby/nativedebug/README.md has details on usage.
The basic APIs for the CRL Reloading features.
This adds external types to represent CRL Providers, CRLs, and
CertificateInfo.
Internally we will use `CrlImpl` - this layer is needed to hide OpenSSL
details from the user.
GRFC - https://github.com/grpc/proposal/pull/382
Things Done
* Add external API for `CrlProvider`, `Crl`, `CertInfo` (`CertInfo` is
used during CRL lookup rather than passing the entire certificate).
* Add code paths in `ssl_transport_security` to utilize CRL providers
* Add `StaticCrlProvider`
* Refactor `crl_ssl_transport_security_test.cc` so it is more extensible
and can be used with providers
@stanley-cheung noticed a bug where CSM labels were not being added on
metrics if the peer was not also CSM Observability enabled.
This PR fixes the behavior to add in the local labels in this case, as
well as add the remote workload type label with the value of unknown.
1. Changes the resource retention period to 2 days for all resources
(previously 7 days for TD resources, 6 hours for k8s). This solved a
problem with k8s resources being stuck because corresponding TD
resources weren't deleted.
2. Resume on namespace cleanup failures
3. Add secondary lb cluster cleanup logic
4. Modularize `grpc_xds_resource_cleanup.sh`
5. Make `KubernetesNamespace`'s methods `pretty_format_status` and
`pretty_format_metadata` public
6. `pretty_format_status`: also print resource kind, creation and
deletion requested dates
ref b/259724370, cl/517235715
We're seeing too many debug headers, change it to only log header in
case of error.
<!--
If you know who should review your pull request, please assign it to
that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the
appropriate
lang label.
-->
Instead of initializing the vector to N elements whose values are all
null, we should be reserving N elements before populating the vector.
This bug didn't actually break anything, but it did cause us to allocate
twice the amount of space that we should have.