Building out a new framing layer for chttp2.
The central idea here is to have the framing layer be solely responsible
for serialization of frames, and their deserialization - the framing
layer can reject frames that have invalid syntax - but the enacting of
what that frame means is left to a higher layer.
This class will become foundational for the promise conversion of chttp2
- by eliminating action from the parsing of frames we can reuse this
sensitive code.
Right now the new layer is inactive - there's a test that exercises it
relatively well, and not much more. In the next PRs I'll add an
experiments to enable using this layer or the existing code in the
writing and reading paths.
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
This is the initial implementation of the chaotic-good client transport
write path. There will be a follow-up PR to fulfill the read path.
<!--
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.
-->
Previously black wouldn't install, as it required newer `packaging`
package.
This fixes `pip install -r requirements-dev.txt`. In addition, `black`
in dev dependencies file is changed to `black[d]`, which bundles
`blackd` binary (["black as a
server"](https://black.readthedocs.io/en/stable/usage_and_configuration/black_as_a_server.html)).
The `work_stealing` experiment on its own is not very valuable, so let's
delete it and save CI resources. We have a benchmark for
`GRPC_EXPERIMENTS=event_engine_listener,work_stealing`, which is really
what we care about right now.
Fixes an issue when an active context selected automatically picked up
as context for `secondary_k8s_api_manager`.
This was introducing an error in GAMMA Baseline PoC
```
sys:1: ResourceWarning: unclosed <ssl.SSLSocket fd=4, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('100.71.2.143', 56723), raddr=('35.199.174.232', 443)>
```
Here's how the secondary context is incorrectly falls back to the
default context when `--secondary_kube_context` is not set:
```
k8s.py:142] Using kubernetes context "gke_grpc-testing_us-central1-a_psm-interop-security", active host: https://35.202.85.90
k8s.py:142] Using kubernetes context "None", active host: https://35.202.85.90
```
- Add Github Action to conditionally run PSM Interop unit tests:
- Only run when changes are detected in
`tools/run_tests/xds_k8s_test_driver` or any of the proto files used by
the driver
- Only run against PRs and pushes to `master`, `v1.*.*` branches
- Runs using `python3.9` and `python3.10`
- Ready to be added to the list of required GitHub checks
- Add `tools/run_tests/xds_k8s_test_driver/tests/unit/__main__.py` test
loader that recursively discovers all unit tests in
`tools/run_tests/xds_k8s_test_driver/tests/unit`
- Add basic coverage for `XdsTestClient` and `XdsTestServer` to verify
the test loader picks up all folders
Related:
- First unit tests without automated CI added in #34097
The tests are skipped incorrectly because `config.server_lang` is
incorrectly compared with the string value "java", instead of
`skips.Lang.JAVA`.
This has been broken since #26998.
```
xds_url_map_testcase.py:372] ----- Testing TestTimeoutInRouteRule -----
xds_url_map_testcase.py:373] Logs timezone: UTC
skips.py:121] Skipping TestConfig(client_lang='java', server_lang='java', version='v1.57.x')
[ SKIPPED ] setUpClass (timeout_test.TestTimeoutInRouteRule)
xds_url_map_testcase.py:372] ----- Testing TestTimeoutInApplication -----
xds_url_map_testcase.py:373] Logs timezone: UTC
skips.py:121] Skipping TestConfig(client_lang='java', server_lang='java', version='v1.57.x')
[ SKIPPED ] setUpClass (timeout_test.TestTimeoutInApplication)
```
This is to make sure upgrading packaging module won't break our logic on
version-based version skipping.
This also fixes a small issue with `dev-` prefix - it should only be
allowed on the left side of the comparison.
Context: packaging module needs to be upgraded to be compatible with
`blackd`.
Update from gtcooke94:
This PR adds support to build gRPC and it's tests with OpenSSL3. There were some
hiccups with tests as the tests with openssl haven't been built or exercised in a
few months, so they needed some work to fix.
Right now I expect all test files to pass except the following:
- h2_ssl_cert_test
- ssl_transport_security_utils_test
I confirmed locally that these tests fail with OpenSSL 1.1.1 as well,
thus we are at least not introducing regressions. Thus, I've added compiler directives around these tests so they only build when using BoringSSL.
---------
Co-authored-by: Gregory Cooke <gregorycooke@google.com>
Co-authored-by: Esun Kim <veblush@google.com>
- add debug-only `WorkSerializer::IsRunningInWorkSerializer()` method
and use it in client_channel to verify that subchannels are destroyed in
the `WorkSerializer`
- note: this mechanism uses `std:🧵:id`, so I had to exclude
work_serializer.cc from the core_banned_constructs check
- fix `WorkSerializer::Run()` to unref the callback before releasing
ownership of the `WorkSerializer`, so that any refs captured by the
`std::function<>` will be released before releasing ownership
- fix the WRR timer callback to hop into the `WorkSerializer` to release
its ref to the picker, since that transitively releases refs to
subchannels
- fix subchannel connectivity state notifications to unref the watcher
inside the `WorkSerializer`, since the watcher often transitively holds
refs to subchannels
Basically run each of the subtests (buildtest, distribtest_cpp,
distribtest_python) as a separate bazel target.
- currently the bazel distribtest are the slowest targets in
grpc_bazel_rbe_nonbazel
- the shards are basically independent tests anyway
- when split into multiple targets, they each get a separate target log
so it's easier debug issues since there isn't multiple bazel invocations
in each log.
Currently the bazel invocations provide a link back to the original
kokoro jobs for resultstore and sponge UIs.
This adds another back link to Fusion UI, which has the advantage of
- being able to navigate to the kokoro job overview
- there is a button to trigger a new build (in case the job needs to be
re-run).
- Upgrade bazel
- Reduce the number of places where bazel version needs to be upgraded
in future.
- also make sure the list of bazel versions to test by bazelified tests
is loaded from supported_versions.txt (it was hardcoded before).
- ~~Try upgrading windows RBE build to bazel 6.3.2 as well.~~
The core idea:
- the source of truth for supported bazel versions is in
`bazel/supported_versions.txt`
- the first version listed in `bazel/supported_versions.txt` is
considered to be the "primary" bazel version and is going to be used in
most places thoroughout the repo.
- use templates to include the primary bazel version in testing
dockerfiles and in a newly introduced `.bazelversion` files (which gets
loaded by our existing `tools/bazel` wrapper).
~~Supersedes https://github.com/grpc/grpc/pull/33880~~
Proposed alternative to https://github.com/grpc/grpc/pull/34024.
This version has a simpler, faster busy-count implementation based on a
sharded set of atomic counts: fast increment/decrement operations,
relatively slower summation of total counts (which need to happen much
less frequently).
WRR is showing a very high CPU cost relative to previous solutions, and
it's unclear why this is.
Add two metrics that should help us see the shape of the subchannel sets
that are being passed to high cost systems in order to confirm/deny
theories.
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Needs https://github.com/grpc/grpc/pull/34035 to be merged first. With
newer ccache in available in the gcc12 test image, the build is now
faster so we can reenable the test.
This should get the benchmarks running again. The dotnet benchmark is
broken (unclear if it's still necessary), and the grpc-go benchmark
build currently fails. The go benchmark should be re-enabled when the
dockerfiles are fixed. The rest of the dotnet benchmark configuration /
artifacts should be deleted or fixed as well. @jtattermusch
Based on https://github.com/grpc/grpc/pull/34033
Bunch of cleanup and rebuilding many docker images from scratch
- consolidate the workaround for "dubious ownership" issue reported by
git. Other team members have run into this recently and used similar but
not identical workarounds so some cleanup is due.
- rebuilding many images increases the chance that we fix the "dubious
ownership" git issue early on rather than later on in the one-at-a-time
fashion in the future (and the former will prevent many teammembers from
wasting time on this weird issue).
- Newer version of ccache is needed for some portability tests to be
able to benefit from caching (e.g. the GCC 12 portability test to get
benefits of local disk caching) - this is a prerequisite for reenabling
the bazelified gcc12 portability test.
- upgrade node interop images to debian:11 (since debian jessie is long
past EOL).
<!--
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.
-->
- Extract build metadata for some external dependencies from bazel
build. This is achieved by letting extract_metadata_from_bazel_xml.py
analyze some external libraries and sources. The logic is basically the
same as for internal libraries, I only needed to teach
extract_metadata_from_bazel_xml.py which external libraries it is
allowed to analyze.
* currently, the list of source files is automatically determined for
`z`, `upb`, `re2` and `gtest` dependencies (at least for the case where
we're building in "embedded" mode - e.g. mostly native extensions for
python, php, ruby etc. - cmake has the ability to replace some of these
dependencies by actual cmake dependency.)
- Eliminate the need for manually written gen_build_yaml.py for some
dependencies.
- Make the info on target dependencies in build_autogenerated.yaml more
accurate and complete. Until now, there were some depdendencies that
were allowed to show up in build_autogenerated.yaml and some that were
being skipped. This made generating the CMakeLists.txt and Makefile
quite confusing (since some dependencies are being explicitly mentioned
and some had to be assumed by the build system).
- Overhaul the Makefile
* the Makefile is currently only used internally (e.g. for ruby and PHP
builds)
* until now, the makefile wasn't really using the info about which
targets depend on what libraries, but it was effectively hardcoding the
depedendency data (by magically "knowing" what is the list of all the
stuff that e.g. "grpc" depends on).
* After the overhaul, the Makefile.template now actually looks at the
library dependencies and uses them when generating the makefile. This
gives a more correct and easier to maintain makefile.
* since csharp is no longer on the master branch, remove all mentions of
"csharp" targets in the Makefile.
Other notable changes:
- make extract_metadata_from_bazel_xml.py capable of resolving workspace
bind() rules (so that it knows the real name of the target that is
referred to as e.g. `//external:xyz`)
TODO:
- [DONE] ~~pkgconfig C++ distribtest~~
- [DONE} ~~update third_party/README to reflect changes in how some deps
get updated now.~~
Planned followups:
- cleanup naming of some targets in build metadata and buildsystem
templates: libssl vs boringssl, ares vs cares etc.
- further cleanup of Makefile
- further cleanup of CMakeLists.txt
- remote the need from manually hardcoding extra metadata for targets in
build_autogenerated.yaml. Either add logic that determines the
properties of targets automatically, or use metadata from bazel BUILD.
Reintroduce https://github.com/grpc/grpc/pull/33959.
I added a fix for the python arm64 build (which is the reason why the
change has been reverted earlier).
Our current implementation of Join, TryJoin leverage some complicated
template stuff to work, which makes them hard to maintain. I've been
thinking about ways to simplify that for some time and had something
like this in mind - using a code generator that's at least a little more
understandable to code generate most of the complexity into a file that
is checkable.
Concurrently - I have a cool optimization in mind - but it requires that
we can move promises after polling, which is a contract change. I'm
going to work through the set of primitives we have in the coming weeks
and change that contract to enable the optimization.
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
The types `google::api::expr::v1alpha1` are available in
`"@com_google_googleapis//google/api/expr/v1alpha1:expr_proto"` and not
`"google_type_expr_upb"`
<!--
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.
-->
Disables the warning produced by kubernetes/client/rest.py calling the
deprecated `urllib3.response.HTTPResponse.getheaders`,
`urllib3.response.HTTPResponse.getheader` methods:
```
venv-test/lib/python3.9/site-packages/kubernetes/client/rest.py:44: DeprecationWarning: HTTPResponse.getheaders() is deprecated and will be removed in urllib3 v2.1.0. Instead access HTTPResponse.headers directly.
return self.urllib3_response.getheaders()
```
This issue introduced by openapi-generator, and solved in `v6.4.0`. To
fix the issue properly, kubernetes/python folks need to regenerate the
library using newer openapi-generator. The most recent release `v27.2.0`
still used openapi-generator
[`v4.3.0`](https://github.com/kubernetes-client/python/blob/v27.2.0/kubernetes/.openapi-generator/VERSION).
Since they release two times a year, and the 2 major version difference
of openapi-generator, the fix may take a while.
Created an issue in their repo:
https://github.com/kubernetes-client/python/issues/2101.
Not adding CMake support yet
<!--
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.
-->
Addresses some issues of the initial triage hint PR:
https://github.com/grpc/grpc/pull/33898.
1. Print unhealthy backend name before the health info - previously it
was unclear health status of which backend is dumped
2. Add missing `retry_err.add_note(note)` calls
3. Turn off the highlighter in triager hints, which isn't rendered
properly in the stack trace saved to junit.xml
Sergii Tkachenko
Michael Lumish
Craig Tiller
AJ Heller
nanahpang
Eugene Ostroukhov
Esun Kim
jrandolf
Mark D. Roth
Jan Tattermusch
Xuan Wang
Stanley Cheung
Terry Wilson
Yash Tibrewal
Yijie Ma