More work on the dualstack backend design:
- Change round_robin to delegate to pick_first instead of creating
subchannels directly.
- Change pick_first such that when it is the child of a petiole policy,
it will unconditionally start a health watch.
- Change the client-side health checking code such that if client-side
health checking is not enabled, it will return the subchannel's raw
connectivity state.
- As part of this, we introduce a new endpoint_list library to be used
by petiole policies, which is intended to replace the existing
subchannel_list library. The only policy that will still directly
interact with subchannels is pick_first, so the relevant parts of the
subchannel_list functionality have been copied directly into that
policy. The subchannel_list library will be removed after all petiole
policies are updated to delegate to pick_first.
The address attribute interface was intended to provide a mechanism to
pass attributes separately from channel args, for values that do not
affect subchannel behavior and therefore do not need to be present in
the subchannel key, which does include channel args. However, the
mechanism as currently designed is fairly clunky and is probably not the
direction we will want to go in the long term.
Eventually, we will want some mechanism for registering channel args,
which would provide a cleaner way to indicate that a given channel arg
should not be used in the subchannel key, so that we don't need a
completely different mechanism. For now, this PR is just doing an
interim step, which is to establish a special channel arg key prefix to
indicate that an arg is not needed in the subchannel key.
With some delay, this is a PR for
https://github.com/grpc/grpc/issues/32564 (and previously
https://github.com/grpc/grpc/pull/31791).
I looked into adding a regular `py_test` for this change [as
suggested](https://github.com/grpc/grpc/pull/31791#issuecomment-1423245116)
but I am not aware of any effect that the presence of a .pyi stub file
would have at runtime and where some sort of type-checking in a .py
script would be affected. Stub files are only for use by type checkers &
IDE's. I mean, something like this would work:
```
import helloworld_pb2
py_file = helloworld_pb2.__file__
pyi_file = py_file + 'i’
self.assertTrue(os.path.exists(pyi_file))
```
But that seems really hacky to me. Instead I created a simple rule test
for `py_proto_library` with Bazel Skylib which tests the declared
outputs for an example `py_proto_library` target. Indirectly, this also
tests that the declared output files are actually generated. Please let
me know if this is sufficient.
- Switched from yapf to black
- Reconfigure isort for black
- Resolve black/pylint idiosyncrasies
Note: I used `--experimental-string-processing` because black was
producing "implicit string concatenation", similar to what described
here: https://github.com/psf/black/issues/1837. While currently this
feature is experimental, it will be enabled by default:
https://github.com/psf/black/issues/2188. After running black with the
new string processing so that the generated code merges these `"hello" "
world"` strings concatenations, then I removed
`--experimental-string-processing` for stability, and regenerated the
code again.
To the reviewer: don't even try to open "Files Changed" tab 😄 It's
better to review commit-by-commit, and ignore `run black and isort`.
Revert "Revert "[core] Add support for vsock transport"
(https://github.com/grpc/grpc/pull/33276)"
This reverts commit
c5ade3011a.
And fix the issue which broke the python build.
@markdroth @drfloob please review this PR. Thank you very much.
---------
Co-authored-by: AJ Heller <hork@google.com>
This is a hack to get around an issue on Apple devices caused by the
PosixEventEngine's `poll` poller not supporting fork. This PR disables
the EventEngine poller entirely in Python builds. It will therefore
prevent the release of the EventEngine generally, and prevent any
testing of EventEngine integration with gRPC Python, until the `poll`
poller is fixed.
This is another attempt to add support for vsock in grpc since previous
PRs(#24551, #21745) all closed without merging.
The VSOCK address family facilitates communication between
virtual machines and the host they are running on.
This patch will introduce new scheme: [vsock:cid:port] to
support VSOCK address family.
Fixes#32738.
---------
Signed-off-by: Yadong Qi <yadong.qi@intel.com>
Co-authored-by: AJ Heller <hork@google.com>
Co-authored-by: YadongQi <YadongQi@users.noreply.github.com>
Most of these data structures need to scale a bit like per-cpu, but not
entirely. We can have more than one cpu hit the same instance in most
cases, and probably want to cap out before the hundreds of shards some
platforms have.
<!--
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.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
This test mode tries to create threads wherever it legally can to
maximize the chances of TSAN finding errors in our codebase.
<!--
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.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Reverts grpc/grpc#33088
The internal test is failing because after the change, transitive
dependencies will be included in response too, thus we need add
`empty2_pb2` to response since it's a dependency of
`empty2_extensions_pb2`.
We already did that in OSS test, we need do the same for internal test
case.
The logger uses `absl::FPrintF` to write to stdout. After reading a
number of sources online, I got the impression that `std::fwrite` which
is used by `absl::FPrintF` is atomic so there is no locking required
here.
---------
Co-authored-by: rockspore <rockspore@users.noreply.github.com>
With these, it is actually possible to have typed client stubs where the
return type is correctly inferred.
It's only for the non-streaming calls, because there is
`RequestIterableType` for the streaming ones (but it's just Any with
extra steps and would require much more work).
---------
Co-authored-by: Xuan Wang <xuanwn@google.com>
Reverts grpc/grpc#32909
It's breaking some internal test
(//net/grpc/python:internal_tests/unit/_default_reflection_test), revert
for now to investigate.
This PR implements a work-stealing thread pool for use inside
EventEngine implementations. Because of historical risks here, I've
guarded the new implementation behind an experiment flag:
`GRPC_EXPERIMENTS=work_stealing`. Current default behavior is the
original thread pool implementation.
Benchmarks look very promising:
```
bazel test \
--test_timeout=300 \
--config=opt -c opt \
--test_output=streamed \
--test_arg='--benchmark_format=csv' \
--test_arg='--benchmark_min_time=0.15' \
--test_arg='--benchmark_filter=_FanOut' \
--test_arg='--benchmark_repetitions=15' \
--test_arg='--benchmark_report_aggregates_only=true' \
test/cpp/microbenchmarks:bm_thread_pool
```
2023-05-04: `bm_thread_pool` benchmark results on my local machine (64
core ThreadRipper PRO 3995WX, 256GB memory), comparing this PR to
master:
2023-05-04: `bm_thread_pool` benchmark results in the Linux RBE
environment (unsure of machine configuration, likely small), comparing
this PR to master.
---------
Co-authored-by: drfloob <drfloob@users.noreply.github.com>
Reverts grpc/grpc#32924. This breaks the build again, unfortunately.
From `test/core/event_engine/cf:cf_engine_test`:
```
error: module .../grpc/test/core/event_engine/cf:cf_engine_test does not depend on a module exporting 'grpc/support/port_platform.h'
```
@sampajano I recommend looking into CI tests to catch iOS problems
before merging. We can enable EventEngine experiments in the CI
generally once this PR lands, but this broken test is not one of those
experiments. A normal build should have caught this.
cc @HannahShiSFB
1. `GrpcAuthorizationEngine` creates the logger from the given config in
its ctor.
2. `Evaluate()` invokes audit logging when needed.
---------
Co-authored-by: rockspore <rockspore@users.noreply.github.com>
Fix: https://github.com/grpc/grpc/issues/18075
From comments in https://github.com/grpc/grpc/issues/18075, `CPython`
reinitialize the `GIL` after `pthread_atfork` child handler, thus we
shouldn't use any `GIL` related functions in child handler which is what
we're currently doing, this PR uses `os.register_at_fork` to replace
`pthread_atfork` to prevent any undesired bevahior.
This also seems to fixes a thread hanging issue cased by changes in
core: https://github.com/grpc/grpc/pull/32869
### Testing:
* Passed existing fork tests. (Note that due to some issues in `Bazel`,
this change was not verified by `Bazel runs_per_test`).
* Tested by patch the core PR, was able to fix Python fork tests:
https://github.com/grpc/grpc/pull/32933
Third-party loggers will be added in subsequent PRs once the logger
factory APIs are available to validate the configs here.
This registry is used in `xds_http_rbac_filter.cc` to generate service
config json.
This paves the way for making pick_first the universal leaf policy (see
#32692), which will be needed for the dualstack design. That change will
require changing pick_first to see both the raw connectivity state and
the health-checking connectivity state of a subchannel, so that we can
enable health checking when pick_first is used underneath round_robin
without actually changing the pick_first connectivity logic (currently,
pick_first always disables health checking). To make it possible to do
that, this PR moves the health checking code out of the subchannel and
into a separate API using the same data-watcher mechanism that was added
for ORCA OOB calls.
The PR also creates a separate BUILD target for:
- chttp2 context list
- iomgr buffer_list
- iomgr internal errqueue
This would allow the context list to be included as standalone
dependencies for EventEngine implementations.
<!--
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.
-->
@sampajano
The very non-trivial upgrade of third_party/protobuf to 22.x
This PR strives to be as small as possible and many changes that were
compatible with protobuf 21.x and didn't have to be merged atomically
with the upgrade were already merged.
Due to the complexity of the upgrade, this PR wasn't created
automatically by a tool, but manually. Subsequent upgraded of
third_party/protobuf with our OSS release script should work again once
this change is merged.
This is best reviewed commit-by-commit, I tried to group changes in
logical areas.
Notable changes:
- the upgrade of third_party/protobuf submodule, the bazel protobuf
dependency itself
- upgrade of UPB dependency to 22.x (in the past, we used to always
upgrade upb to "main", but upb now has release branch as well). UPB
needs to be upgraded atomically with protobuf since there's a de-facto
circular dependency (new protobuf depends on new upb, which depends on
new protobuf for codegen).
- some protobuf and upb bazel rules are now aliases, so `
extract_metadata_from_bazel_xml.py` and `gen_upb_api_from_bazel_xml.py`
had to be modified to be able to follow aliases and reach the actual
aliased targets.
- some protobuf public headers were renamed, so especially
`src/compiler` needed to be updated to use the new headers.
- protobuf and upb now both depend on utf8_range project, so since we
bundle upb with grpc in some languages, we now have to bundle utf8_range
as well (hence changes in build for python, PHP, objC, cmake etc).
- protoc now depends on absl and utf8_range (previously protobuf had
absl dependency, but not for the codegen part), so python's
make_grpcio_tools.py required partial rewrite to be able to handle those
dependencies in the grpcio_tools build.
- many updates and fixes required for C++ distribtests (currently they
all pass, but we'll probably need to follow up, make protobuf's and
grpc's handling of dependencies more aligned and revisit the
distribtests)
- bunch of other changes mostly due to overhaul of protobuf's and upb's
internal build layout.
TODOs:
- [DONE] make sure IWYU and clang_tidy_code pass
- create a list of followups (e.g. work to reenable the few tests I had
to disable and to remove workaround I had to use)
- [DONE in cl/523706129] figure out problem(s) with internal import
---------
Co-authored-by: Craig Tiller <ctiller@google.com>
Various fixes that can be merged in advance, to simplify the 22.x
upgrade.
- Use NOMINMAX for grpcio_tools build to avoid build failure when absl
is added as a dependency. Fixes
https://github.com/grpc/grpc/issues/32779
- in spawn_patch command file, put arguments on multiple lines to avoid
lines going over the limit for large link commands: Fixes
https://github.com/grpc/grpc/issues/32795
- Use `python/generator.h` instead of the deprecated
`python/python_generator.h` in grpcio_tools.
### Description
When invoking a RPC, sometimes operations in core will fail which result
in an `ExecuteBatchError`, currently we [simply raise this
error](https://github.com/grpc/grpc/blob/master/src/python/grpcio/grpc/_cython/_cygrpc/aio/server.pyx.pxi#L705),
and left this Exception unhandled in event loop with error message `Task
exception was never retrieved`.
### Changes included
This PR changes the behavior to log the `ExecuteBatchError` instead of
raising it.
This shouldn't change existing behavior because even without this
change, the exception will simply left on event loop.
### Testing
Tested by manually `set_expection` in [this
future](https://github.com/grpc/grpc/blob/master/src/python/grpcio/grpc/_cython/_cygrpc/aio/callback_common.pyx.pxi#L84),
verified that error was logged correctly and no exception was left in
event loop, sample log:
```
ERROR:grpc._cython.cygrpc:ExecuteBatchError raised in core by servicer method [/grpc.testing.TestService/UnaryCall]
Traceback (most recent call last):
File "src/python/grpcio/grpc/_cython/_cygrpc/aio/server.pyx.pxi", line 682, in _cython.cygrpc._handle_exceptions
File "src/python/grpcio/grpc/_cython/_cygrpc/aio/server.pyx.pxi", line 802, in _handle_rpc
File "src/python/grpcio/grpc/_cython/_cygrpc/aio/server.pyx.pxi", line 547, in _handle_unary_unary_rpc
File "src/python/grpcio/grpc/_cython/_cygrpc/aio/server.pyx.pxi", line 452, in _finish_handler_with_unary_response
File "src/python/grpcio/grpc/_cython/_cygrpc/aio/callback_common.pyx.pxi", line 106, in execute_wrong_batch
_cython.cygrpc.ExecuteBatchError: Failed "execute_batch_including_SendInitialMetadataOperation": (<_cython.cygrpc.SendInitialMetadataOperation object at 0x7fa86d936ca0>, <_cython.cygrpc.SendMessageOperation object at 0x7fa86cf48cb0>, <_cython.cygrpc.SendStatusFromServerOperation object at 0x7fa86cf4c110>)
```
Fix python type hints for async iteration on `UnaryStreamCall` and
`StreamStreamCall`. Those classes _are_ `AsyncIterable`s and `__aiter__`
returns an `AsyncIterator`.
<!--
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.
-->
This PR also centralizes the client channel resolver selection. Resolver
selection is still done using the plugin system, but when the Ares and
native client channel resolvers go away, we can consider bootstrapping
this differently.
<!--
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.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
(hopefully last try)
Add new channel arg GRPC_ARG_ABSOLUTE_MAX_METADATA_SIZE as hard limit
for metadata. Change GRPC_ARG_MAX_METADATA_SIZE to be a soft limit.
Behavior is as follows:
Hard limit
(1) if hard limit is explicitly set, this will be used.
(2) if hard limit is not explicitly set, maximum of default and soft
limit * 1.25 (if soft limit is set) will be used.
Soft limit
(1) if soft limit is explicitly set, this will be used.
(2) if soft limit is not explicitly set, maximum of default and hard
limit * 0.8 (if hard limit is set) will be used.
Requests between soft and hard limit will be rejected randomly, requests
above hard limit will be rejected.
<!--
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.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
This is a big rewrite of global config.
It does a few things, all somewhat intertwined:
1. centralize the list of configuration we have to a yaml file that can
be parsed, and code generated from it
2. add an initialization and a reset stage so that config vars can be
centrally accessed very quickly without the need for caching them
3. makes the syntax more C++ like (less macros!)
4. (optionally) adds absl flags to the OSS build
This first round of changes is intended to keep the system where it is
without major changes. We pick up absl flags to match internal code and
remove one point of deviation - but importantly continue to read from
the environment variables. In doing so we don't force absl flags on our
customers - it's possible to configure grpc without the flags - but
instead allow users that do use absl flags to configure grpc using that
mechanism. Importantly this lets internal customers configure grpc the
same everywhere.
Future changes along this path will be two-fold:
1. Move documentation generation into the code generation step, so that
within the source of truth yaml file we can find all documentation and
data about a configuration knob - eliminating the chance of forgetting
to document something in all the right places.
2. Provide fuzzing over configurations. Currently most config variables
get stashed in static constants across the codebase. To fuzz over these
we'd need a way to reset those cached values between fuzzing rounds,
something that is terrifically difficult right now, but with these
changes should simply be a reset on `ConfigVars`.
<!--
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.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Prior to this change, we invoke aycnio interceptors by converting them
to `iterator` first, which means we can only call them once before
they're exhausted.
This PR changes the implementation to use `list`, thus the interceptors
can be called multiple times.
<!--
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.
-->
---------
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
This change mostly aims to get OpenCensus to use the new
ServerCallTracer interface. Note that the interfaces nor the code are in
their final states. There are a bunch of moving pieces, but I thought
this might be a nice mid-step to check-in and make sure that our
internal traces can also work with these changes.
Overall changes -
1) call_tracer.h shows what the hierarchy of new call tracer interfaces
looks like. Open to renaming suggestions.
2) Moved most of the common interface between `CallAttemptTracer` and
`ServerCallTracer` into a common `CallTracerInterface`. We should be
able to eventually move `RecordReceivedTrailingMetadata` and `RecordEnd`
as well to these common interfaces, but it requires some additional
work.
3) The compression filter is now responsible for recording the recv and
send messages for both the subchannel call and the server, and adds in
ability to record compressed and decompressed messages as well.
4) The OpenCensus server filter now uses the new `ServerCallTracer`
interface, and so doesn't need to be a filter anymore.
5) A new ServerCallTracerFilter was added. Ideally, we should be able to
move it to the current connected filter, but it is in a bit of an
interesting state right now, so I would prefer making those changes in a
separate PR with Craig's eyes on it.
6) A new context element `GRPC_CONTEXT_CALL_TRACER_ANNOTATION_INTERFACE`
was created that replaces the old `GRPC_CONTEXT_CALL_TRACER`, and the
new `GRPC_CONTEXT_CALL_TRACER` is mainly to pass the `CallAttemptTracer`
down the stack. This should go away in the new promise-based world.
<!--
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.
-->
<!-- Reviewable:start -->
- - -
This change is [<img src="https://reviewable.io/review_button.svg"
height="34" align="absmiddle"
alt="Reviewable"/>](https://reviewable.io/reviews/grpc/grpc/32618)
<!-- Reviewable:end -->
<!--
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.
-->
It looks like nobody ever created ALTS redentials from Python with a
list of accepted service accounts before.
Simple reproduction:
```
import grpc
grpc.alts_channel_credentials(None) # works
grpc.alts_channel_credentials(['foo']) # fails
```
Without this change, generates this error:
```
[...]
File "src/python/grpcio/grpc/_cython/_cygrpc/credentials.pyx.pxi", line 414, in _cython.cygrpc.channel_credentials_alts
File "src/python/grpcio/grpc/_cython/_cygrpc/credentials.pyx.pxi", line 403, in _cython.cygrpc.ALTSChannelCredentials.__cinit__
TypeError: expected bytes, str found
```
(And the error cannot be worked around by the caller by passing a bytes
object from the Python side: you still get the same error.)
Upgrade boringssl to the latest "master-with-bazel"
- use the `'USE_HEADERMAP' => 'NO'` fix for ObjC
- update the key for asm optimizations on mac/apple in python's setup.py
This PR depends on monterey fixes here:
https://github.com/grpc/grpc/pull/32493 and the boringssl's build
simplification
https://boringssl-review.googlesource.com/c/boringssl/+/56465.
---------
Co-authored-by: Hannah Shi <hannahshisfb@gmail.com>
Internal Windows builds will catch issues that we cannot yet catch in
OSS. This will be mostly remedied once we have clang-cl in our CI (See a
rough roadmap in https://github.com/grpc/grpc/pull/32448). For now, this
PR identifies folders where most Windows-specific code is developed, and
requires cherrypicks for PRs that touch anything inside those folders.
This PR also refactors gpr and gprpp source files to better isolate all
platform-specific code ~the Windows-only code~. ~I will reorganize the
other platform-specific files using this structure if there are no
objections.~
This subset of folders covers about half of the `#ifdef GPR_WINDOWS`
usages in gRPC, but nearly all of the actively-developed Windows code
locations.
This version broke backward compatibility in `plugin_pb2.py`, which is
presumably a relatively minor regression, since we have not yet heard
any complaints about it. This PR:
- Excludes `4.22.0` from installation
- _Includes_ protobuf pre-releases into testing so this can be caught
more quickly in the future.
When bad prereleases are caught, we can exclude them from testing in a
similar manner to this PR. We may eventually want to invest into a
system where we can define these bad versions centrally.
This code is not plumbed through yet, but it provides the core
infrastructure needed to detect the proper GCP environment resources
needed to set up the labels/attributes/resources for stats, tracing and
logging.
Details on how the various environment resources are setup has been
derived by looking at java's cloud logging library and OpenTelemetry's
future plans. (Could be better explained in an offline review since some
links are internal).
Requesting @veblush for a full review and @markdroth for a structural
review.
<!--
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.
-->
* Revert "Revert "Revert "Revert "server: introduce ServerMetricRecorder API and move per-call reporting from a C++ interceptor to a C-core filter (#32106)" (#32272)" (#32279)" (#32293)"
This reverts commit 1f960697c5.
* Do not create CallMetricRecorder if call is null.
* Revert "Revert "server: introduce ServerMetricRecorder API and move per-call reporting from a C++ interceptor to a C-core filter (#32106)" (#32272)"
This reverts commit deb1e25543.
* Fix by caching call metric recording stuff in async request
PR #32106 caused msan errors in some tests while de-referencing the
server object where async calls are active after the server is
destroyed. Instead cache the ServerMetricRecorder pointer.
* copyright headers fixed
* clang fixes.
* WIP. A seemingly properly failing test
* WIP. Pre-fork handlers now work
* Roughly working.
* Clean up
* Clean up more
* Add to CI
* Format
* Ugh. Remove swap file
* And another
* clean up
* Add copyright
* Format
* Remove another debug line
* Add stub forkable methods
* Remove use of 3.9+ function
* Remove unintentional double copyright
* drfloob review comments
* Only hold lock during Close once
* Create separate job for fork test
* Bump up gdb timeout
* Format
* [EventEngine] Add invalid handle types to the public API
* Automated change: Fix sanity tests
* add definition for static constexpr members
* sanitize
Co-authored-by: drfloob <drfloob@users.noreply.github.com>
* WRR: port StaticStrideScheduler to OSS
* WIP
* Automated change: Fix sanity tests
* fix build
* remove unused aliases
* fix another type mismatch
* remove unnecessary include
* move benchmarks to their own file, and don't run it on windows
* Automated change: Fix sanity tests
* add OOB reporting
* generate_projects
* clang-format
* add config parser test
* clang-tidy and minimize lock contention
* add config defaults
* add oob_reporting_period config field and add basic test
* Automated change: Fix sanity tests
* fix test
* change test to use basic RR
* WIP: started exposing peer address to LB policy API
* first WRR test passing!
* small cleanup
* port RR fix to WRR
* test helper refactoring
* more test helper refactoring
* WIP: trying to fix test to have the right weights
* more WIP -- need to make pickers DualRefCounted
* fix timer ref handling and get tests working
* clang-format
* iwyu and generate_projects
* fix build
* add test for OOB reporting
* keep only READY subchannels in the picker
* add file missed in a previous commit
* fix sanity
* iwyu
* add weight expiration period
* add tests for weight update period and OOB reporting period
* Automated change: Fix sanity tests
* lower bound for timer interval
* consistently apply grpc_test_slowdown_factor()
* cache time in test
* add blackout_period tests
* avoid some unnecessary copies
* clang-format
* add field to config test
* simplify orca watcher tracking
* attempt to fix build
* iwyu
* generate_projects
* update xds proto dependency
* add xDS LB policy entry to registry
* add "_experimental" suffix to policy name
* update LB policy name and remove debug log
* add env var protection
* generate_projects
* gen_upb_api
* WRR: update tests to cover qps plumbing
* WIP
* Automated change: Fix sanity tests
* more WIP
* basic WRR e2e test working
* add OOB test
* add xDS WRR e2e test
* clang-format
* fix sanity
* ignore duplicate addresses
* Automated change: Fix sanity tests
* add new tracer to doc/environment_variables.md
* retain scheduler state across pickers
* Automated change: Fix sanity tests
* use separate mutexes for scheduler and timer
* sort addresses to avoid index churn
* remove fetch_sub for wrap around in RR case
Co-authored-by: markdroth <markdroth@users.noreply.github.com>
* WRR: port StaticStrideScheduler to OSS
* WIP
* Automated change: Fix sanity tests
* fix build
* remove unused aliases
* fix another type mismatch
* remove unnecessary include
* move benchmarks to their own file, and don't run it on windows
* Automated change: Fix sanity tests
* add OOB reporting
* generate_projects
* clang-format
* add config parser test
* clang-tidy and minimize lock contention
* add config defaults
* add oob_reporting_period config field and add basic test
* Automated change: Fix sanity tests
* fix test
* change test to use basic RR
* WIP: started exposing peer address to LB policy API
* first WRR test passing!
* small cleanup
* port RR fix to WRR
* test helper refactoring
* more test helper refactoring
* WIP: trying to fix test to have the right weights
* more WIP -- need to make pickers DualRefCounted
* fix timer ref handling and get tests working
* clang-format
* iwyu and generate_projects
* fix build
* add test for OOB reporting
* keep only READY subchannels in the picker
* add file missed in a previous commit
* fix sanity
* iwyu
* add weight expiration period
* add tests for weight update period and OOB reporting period
* Automated change: Fix sanity tests
* lower bound for timer interval
* consistently apply grpc_test_slowdown_factor()
* cache time in test
* add blackout_period tests
* avoid some unnecessary copies
* clang-format
* add field to config test
* simplify orca watcher tracking
* attempt to fix build
* iwyu
* generate_projects
* add "_experimental" suffix to policy name
* WRR: update tests to cover qps plumbing
* WIP
* more WIP
* basic WRR e2e test working
* add OOB test
* fix sanity
* ignore duplicate addresses
* Automated change: Fix sanity tests
* add new tracer to doc/environment_variables.md
* retain scheduler state across pickers
* Automated change: Fix sanity tests
* use separate mutexes for scheduler and timer
* sort addresses to avoid index churn
* remove fetch_sub for wrap around in RR case
Co-authored-by: markdroth <markdroth@users.noreply.github.com>
* Improve logging while calling core API
* Use NULL for _core_error_string
* Don't raise inside Exception hanlder
* Changes based on comment
* Remove unused enum
* Decode and add qoutes for grpc_call_error_string.
* Revert "Revert "Add support for systemd socket activation (#30485)" (#31617)"
This reverts commit 867dc6cae2.
* Add checks to unix tests
* Ran generate_projects.sh and fixed styling in test
* Fix variable in unit test
* Use reinterpret_cast in test
* Rebase and fix sanity failures
This is motivated by #31227 which I also encountered recently. I'm not familiar with the call lifecycle details, but it seems clear that there is some kind of race condition between user-initiated cancellations and other completion status updates.
If a user cancels a call before being notified of its completion, I believe it should be considered cancelled unconditionally, regardless of whether the call actually completed successfully before its cancellation status could be set.
So it seems better to re-raise explicitly rather than using self._raise_for_status(), which is a no-op in this race condition and can lead to an UnboundedLocalError.
* rename task-handle mutex
* rename TimerClosure
* tmp commit, building, not tested
* add test for client connection to a non-listening port
* fix posix EE tests
* re-fix windows test suite after posix compatibility
* (unfinished, backup): passing the suite's NonExistingListener client test
* remove redundant windows client test
* integrate IOCP worker loop
* initial commit of echo test tool; fixes
* move echo client to test_suite/tools; I do not yet like the setup, it's about time for a macro that generates all useful test/tool targets
* cleanup
* add --target arg to echo_client. requires URI
* Automated change: Fix sanity tests
* build fixes
* fix
* fix
* reviewer feedback
* warning fix
* delete logic on cancellation
* fix
* cancel connect deadlock; improved template code
* fix build failure with linux environments building windows targets
* fix
* fix
* no ++ for atomics
* remove the test changes, to be landed separately
* Automated change: Fix sanity tests
* remnants
Co-authored-by: drfloob <drfloob@users.noreply.github.com>
* Revert "Revert "xDS stateful session affinity: add config plumbing (#31827)" (#31873)"
This reverts commit 4f15d3dcf9.
* fix build for compilers too dumb to recognize the full set of enum values
* Add type annotation to _channel.py and _server.py
* Undo changes in init.py
* changes based on comments.
* format changes
* fix pytype errors
* Changes based on comments
* Fix circular import
* fix pytype and pylint
* Fix format issues
* Fix some cython type issues
* Change ManagedCallType to IntegratedCallFactory
* fix format
* undo some mistakes
Add a check that SEND_STATUS_FROM_SERVER and RECV_MESSAGE are not in the same batch.
This is necessary pre-work for #31204 and implements part of grpc/proposal#336.
Also eliminates fling instead of updating it:
My expectation is nobody has looked at this corner in many years
It's not a benchmark we want: concentrating on a microbenchmark that doesn't include a binding layer caused us to favor designs that emphasized a lightweight core at the expense of a expensive bindings. We should consider the whole.
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
* Actually record CPU usage
* Do the client path too
* Actually build the image from the PR
* Revert "Actually build the image from the PR"
This reverts commit 3afa5cf3c3.
* Support importing on Windows
* Declarative JSON parser
* Automated change: Fix sanity tests
* fix
* shrinking stuff a little
* static vtables
* separate fns
* simpler?
* make maps work
* windows fixes
* Automated change: Fix sanity tests
* simplify code
* Automated change: Fix sanity tests
* vtable-test
* dont always create vec/map impls for every type
* comments
* make error consistent
* move method private
* progress
* durations!
* Automated change: Fix sanity tests
* fix
* fix
* fix
* Automated change: Fix sanity tests
* post-load
* Automated change: Fix sanity tests
* document JsonPostLoad() and add static_assert
* don't copy field names, to avoid length limitations
* use absl::Status
* accept either string or number for numeric values
* add test for direct data member of another struct type
* remove unused method
* add support for retaining part of the JSON wirthout processing
* update test for changes in Json::Parse() API
* add absl::optional support
* Automated change: Fix sanity tests
* fix tests, improve error messages, and add overload to parse to existing object
* remove overload of LoadFromJson()
* change special case for Json to instead use Json::Object
* rename resolver_result_parsing to client_channel_service_config
* split up service_config_test into a separate test for each component
* convert client channel service config parser to new API
* fix build
* converted retry global params
* convert retry method-level parsing, but still need to find a way to control parsing via a channel arg
* improve error structure, add missing types, and improve tests
* clang-format
* Automated change: Fix sanity tests
* fix build
* add LoadJsonObjectField(), add LoadFromJson() overload that takes an ErrorList parameter, and add tests for parsing bare top-level types
* fix msan
* Automated change: Fix sanity tests
* fix error message
* Automated change: Fix sanity tests
* fix test
* add ability to disable fields
* plumb in channel args to disable parsing
* Automated change: Fix sanity tests
* use const char* instead of absl::string_view for enable_key
* fix resolver_component_test
* Automated change: Fix sanity tests
* work around mac build problem
* Automated change: Fix sanity tests
* work around gcc6 problem
* Automated change: Fix sanity tests
* fix build
* fix build
* don't use alternative builder in tests
* convert message size service config parser
* convert fault injection service config parser
* rename files
* add specialization for unique_ptr
* avoid moves in client channel service config parser
* avoid moves in retry service config parser, and do some cleanup
* avoid moves in message_size service config parser
* avoid moves for fault injection service config parser, and use internal channel arg
* convert rbac service config parser
* clang-format
* WIP
* convert top-level service config parser
* clang-format
* fix build
* fix rbac service config parser test
* fix signed-ness problem and reversed-conditional bug
* fix unused param
* fix json_string method
* fix max message length defaults
* add copy ctors to appease windows compiler
* fix RLS LB config parser test
* fix name resolution test
* fix build
* work around gmock portability bug
* fix sanity
* add missing build dep
* make RBAC principal and permissions movable, not copyable
* Revert "make RBAC principal and permissions movable, not copyable"
This reverts commit 53315bccc9.
* attempt to simplify HeaderMatcher and StringMatcher parsing
* more bloat reduction in RBAC service config parser
* fix sanity
* add missing build dep
* attempt work-around for MSVC bug
* Revert "attempt work-around for MSVC bug"
This reverts commit e54c89e1e4.
* attempt work-around for Windows build problem
* try another work-around
* fix sanity
* appease clang-tidy
* generate_projects
* attempt to fix Windows build
* more windows fixes
* more windows fix
* yet more windows fixes
* try without noexcept
* remove unnecessary boilerplate
* code review changes
* fix breakage
Co-authored-by: Craig Tiller <craig.tiller@gmail.com>
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Co-authored-by: Craig Tiller <ctiller@google.com>
Co-authored-by: markdroth <markdroth@users.noreply.github.com>
* XdsBootstrap: move two more methods out of the interface
* Automated change: Fix sanity tests
* XdsClient: add unit test
* Automated change: Fix sanity tests
* fix memory leaks
* add helper method
* add unsubscription
* add test for multiple subscriptions
* clang-format
* fix build
* fix flakiness
* add checking for other node fields
* add v2 test
* add response builder
* add test for update from server
* add test for update containing only changed resources
* clang-format
* fix build
* add test for resource not existing upon subscription
* add test for stream closed by server
* add test for multiple watchers for the same resource
* add test for connection failure
* clang-format
* add test for resources wrapped in Resource wrapper message
* add test for resource validation failure
* add test for multiple invalid resources, and fix a case in XdsClient
* add test for validation failure for already-cached resource
* add test for server not resending resources after stream disconnect
* clang-format
* fix XdsClient to report channel errors to newly started watchers
* fix XdsClient to send cached errors/does-not-exists to newly started watchers
* fix watcher to ensure events arrive in the expected order
* fix tests
* clang-format
* add test for multiple resource types
* fix xds_cluster_e2e_test
* Automated change: Fix sanity tests
* cleanup
* add federation tests
* clang-format
* remove now-unnecessary XdsCertificateProviderPluginMapInterface
* code review comments
* simplify XdsResourceType::Decode() API
* XdsClient: add unit tests for XdsClusterResourceType
* add XdsClient with gRPC bootstrap config
* add LB policy tests
* started adding CertificateProvider tests
* update for recent API changes
* fix merge bugs
* xDS resource validation: identify extensions by type_url instead of name
* fix build
* migrate to ValidationErrors
* add xds_common_types_test
* finish TLS tests and add LRS tests
* move ScopedExperimentalEnvVar to its own library and remove redundant e2e tests
* add circuit breaking and outlier detection tests
* add validation to outlier detection LB policy parsing
* clang-format
* Automated change: Fix sanity tests
* fix signedness
* fix sanity
* xDS: implement xds_wrr_locality LB policy and return xDS LB config from XdsClient
* fix unused parameter
* fix sanity
* fix test
* Automated change: Fix sanity tests
* fix aggregate cluster bug
* Automated change: Fix sanity tests
* absl::make_unique -> std::make_unique
* fix sanity
* fix sanity
* iwyu
* iwyu
* update code for XdsResourceTypeImpl changes
Co-authored-by: markdroth <markdroth@users.noreply.github.com>
* begin c++
* Automated change: Fix sanity tests
* progress
* progress
* missing-files
* Automated change: Fix sanity tests
* moved-from-stats
* remove old benchmark cruft, get tests compiling
* iwyu
* Automated change: Fix sanity tests
* fix
* fix
* fixes
* fixes
* add needed constructor
* Automated change: Fix sanity tests
* iwyu
* fix
* fix?
* fix
* fix
* Remove ResetDefaultEventEngine
Now that it is a weak_ptr, there's no need to explicitly reset it. When
the tracked shared_ptr is deleted, the weak_ptr will fail to lock, and a
new default EventEngine will be created.
* forget existing engine with FactoryReset
* add visibility
* fix
Co-authored-by: ctiller <ctiller@users.noreply.github.com>
Co-authored-by: AJ Heller <hork@google.com>
* Re-add compiler protos to grpcio-tools
* Attempt to break
* Actually fail if protobufs don't compile. Who wrote this thing?
* Add an actual test
* yapf
* Whoops. Thought I'd reverted you
* Please the formatter gods
* Support Python 3.11
* Update build images for 3.11
* Whoopsie
* The architecture of this thing is garbage
* Silence ownership warning
* Account for change in git behavior
* Fix directory
* I am in great pain
* Update Windows and arm linux
* Agh
* Clean up
* [cleanup] Remove profiling timers
- nobody has used this system in years
- if we needed it, we'd probably rewrite it at this point to be something more modern
- let's remove it until that need arises
* fix
* fixes
A (currently) pthread_atfork-based fork support mechanism, allowing EventEngines - or any other object that wants to implement the Forkable interface - respond to forks.
* Rename the default EventEngine headers
Small cleanup. This code hasn't been related to factories for a month or
two.
* ensure only one target contains default_event_engine.h
* src + hdr in same target
* include guards
* Update third_party/protobuf
* run tools/distrib/python/make_grpcio_tools.py
* regenerate protos for ruby, php
* update build_handwritten.yaml
* regenerate projects
* Build - Use :well_known_type_protos instead of :well_known_protos
* Fix target
* Update upb
* Update Python for Protobuf 4.21 (#140)
* Update protobuf dependency on grpcio-tools
* Off by one
* Drop python 3.6 support
* Try upgrading pip
* And in the other script
* Try to figure out if we're compatible with abi3
* See what we've already got installed
* Update the requirements.txt file I didn't know existed
* And here too
* See what's installed
* Let's try that again
* Remove
* Try to confirm version
* Let me see the generated code
* Fix non-Bazel test runner
* Work for all test directories
* Regenerate example protos
* Clean up
* Generate .pyi files
* Fix type checking and linting
* Exclude pyi files from isort
* Upgrade to 3.21.4
* Update iwyu to get around messy protobuf IWYU rules
Co-authored-by: Richard Belleville <gnossen@gmail.com>
This is a partial fork of the windows iomgr code - specifically the IOCP and Socket pieces - with some improved architecture and encapsulation. And the start of a WindowsEventEngine.
Once this code is used in a gRPC TCP context, I imagine a few issues will shake out. Also, getting sanitizers set up with MSVC will take a bit of work (see a commit referencing abseil and MSVC bugs to hack around).
I forked the IomgrEventEngine's posix poller interfaces in the hope of negotiating compatibility between the platforms, but the interfaces diverged a fair bit, and I'm doubtful we'll be able to use these "pollers" generically in the same TCP code. Reunification might not happen, and that's probably fine, we'll see how similar the TCP code looks once it's fleshed out.
I also extracted the IomgrEventEngine's timer piece into a separate component, usable by both engines.
* xDS: upgrade the commit of the xDS protos that we're using
* Automated change: Fix sanity tests
Co-authored-by: markdroth <markdroth@users.noreply.github.com>
This change includes:
* adding a cert file path for MacOS
* updating related test to run on MacOS too
* s/linux/supported/ since this now includes more platforms
* regenerating files affected by the name change
* [Aio] Ensure Core channel closes when deallocated
* Keep channel object alive
* Let Multicallable objects hold the reference to Channel
* Make YAPF happy
This code is not compiled by default and has fallen out of sync with the
rest of the codebase. There's a good chance it won't be used, given our
current work to use an iomgr-drive EventEngine instead.
This code will continue to live in git history, should we need to bring
pieces of it back.
* Initial skeleton for outlier detection
* fixing code review comments (modifying child policy)
* Skeleton and all tests passing except for 1
* small code review comments fix
* Adding the parsing of the policy in cds and put it in discovery
mechansim json format
* Parsing outlier detection json policy from parent
* Adding parsing of the updates
* Added Subchannel wrapper and watcher wrapper: and all states pass
through and all tests still pass
* added framework to do eject and uneject
* fixing code review comments
* restore a test
* fixing code review comments
* taking care of code review comments
* removing debug code and rebuild build files
* fixing according to code review comments
* fixing code review comments
* Adding address to subchannel map
* addressing code review comments
* adding call counter
* Refcount SubchannelState (in the map) and store them in Subcahnnel Wrapper
* fixing counterss
* Call counter and tracker skleton added
* Call counter
* addressing code review comments
* addressing code review comments
* Added CallCounter and timer
* fixing sanity; but more importantly: taking out timer temporarly as it
was causing test failures.
* sanity
* fixing according to code review comments
* addressing code review comments
* all algorithms implemented
* addressing code review comment about starting the timer
* protect private vars
* small fix
* Added one more corner case
* fixing EjectionTimer
* Fixing according to code review suggestions.
* fixing according to code reveiw comments
* taking care of code review comments
* fixing sanity issues
* Adding proto to tests
* First test
* Fixing according to code review comments
* Tests all working now
* fixing a crash
* fixing build files
* fixing sanity
* sanity
* Simplifying tests
* merge and update
* format
* sanity and format
* Fixing asan error
* fixing parsing logic and error handling
* 6 more tests done
* Added verifying unejection to tests
* Added all the tests
* fixing according to code review comments
* fixing asan and ubsan
* Fixing tests according to code review comments
* Added both algorithm tests
* added percentage enforcement tests
* fixing tsan error
* keeping debug, but fix warning
* remove debugs
* fixing IWYU and build errors after
* test comments change only but very important
* fixing code review comments
* one more refactorying of util function
* Removed debugs and added one more helper method
* one more logic fix
* Fixing last bit of code review comments and added disable tests
* fixing code review comments
* fixing IWYU
* sanity format
* protecting the feature with environment var:
registering policy and generating policy
* added a todo according to code review comments
* fixing a clang finding at import time
* build fix after synching to latest
Rollforward with TCP connect handshaker again(#29111) after fixing broken internal targets.
The changes needed were just visibility changes to the handshaker and the http_connect_handshaker libraries as they are used internally.
* Move unnecessary absl includes from metadata_batch.h
metadata_batch.h is a widely used header, and so any includes it takes
that are for its internal use can negatively impact our build times.
* missing file
* build
* review feedback
Currently the tcp connect is performed in chttp2_connector before the handshaking is triggered. For
use cases where the application wants to perform business logic before
the tcp connection, this is problematic. By moving the TCP connect into
its own handshaker and registering it by default at the beginning, this
allows applications to add a new handshaker at the beginning allowing
handshaker logic before a TCP connect.
This approach has the advantage of slightly simplifying the logic in
tcp_connect_handshaker and httpcli as tcp_connect/callback can be
removed.
As the TCP connect needs parameters like resolved_addr,
interested_parties, a new struct called connection args is created as a
member for Handshaker Args.
For server handshakers most of the arguments here are not directly
useful, other than the deadline.
Eugene Ostroukhov
