* milestone 1: static build instance, cmake
* on-demand vm per build, and use bazel
* PR cleanup
* pr cleanup: use builtin bazel wrapper
* pr: misc cleanup
* less verbose unzip
* small cleanup of shell scripts and config file
* using rsync for copying workspace is much faster
* simpler way to increase worker disk size
* simplify bazel build
* increase job timeout
* make max instance lifespan setting more obvious
* refactor the exitcode logic
* shutdown AWS instance once possible
* sudo shudown is required
* add useful AWS instance tags
* move aws integration scripts under AWS folder
* adjust scripts
* make sure ssh session closes even if stdout/stderr remains open
* add test scripts for multiple languages
* improvements to the run_remote_test script
* add cfg files for aws kokoro jobs
Co-authored-by: Alexander Midlash <amidlash@google.com>
* List annotations of tests that have running pods and are in Errored state.
* Fix format.
* Use pod owner reference instead of "loadtest" label.
* Delete loadtests that have running pods and are in Errored state.
* Improve jsonpath expressions.
* Add comment.
* List tests but do not delete them.
* Run all kokoro performance tests on dedicated node pools.
Both official and experimental tests run with kokoro are set to run on separate node pools with the suffix "-ci" (drivers-ci, workers-8core-ci, workers-32core-ci), separate from the default pools used for manual runs.
This change sets the deadline of master and experimental kokoro jobs to match their running interval (4 hours and 12 hours), and makes a change so runs initiated by kokoro are marked 'kokoro' for the master job and 'kokoro-test' for the experimental job.
Experiments show that 4 tests running concurrrently (two on 8-core nodes and two on 32-core nodes) is enough to run all tests within two hours with time to spare.
* Removes optional flag -a, allowing it to be changed later to a long-form flag.
* Updates concurrency levels to one more than what each worker node pool can support (each test requires two workers, and there are nine nodes in each pool, so each node can support four tests).
* Employ prebuilt images in continuous build.
This commits updated Kokoro build job to use prebuilt images to
run tests. The loadtest template was generated using
loadtest_template.py.
pip 21.1 released on Apr 24 introduced a regression for python 3.6.1.
The regression was identified on Apr 24, the fix merged on Apr 25.
The fix is expected to be delivered in the 21.1.1 patch.
There's no clear date, when 21.1.1 will be released.
Until then, pin is temporarily pinned to the previous release, 21.0.1.
This commit includes the following changes:
1. A new load test template generator (loadtest_template.py) is added. The template generator combines existing configurations or templates for several languages into a single template that can be used to generate configurations for different languages or combinations of languages.
2. A basic template generated from the example tests in grpc/test-infra (loadtest_template_basic_all_languages.yaml) is added.
3. The load test config generator is updated to use the combined template.
4. An example run consisting of a single test (generated from the combined template) is added and set up to run continuously.
* Add CSDS xDS interop test
* Add CSDS test to the test suite
* Fix a typo
* Address comments
* Improve the logging of each attempt
* Improve Python readability
* Implement C++ Admin Interface API
* Address reviewer's requests
* Remove static asserts for raw pointers
* Make sanity tests happy
* Windows: pacify conflict between ifndef and macros
* Disable admin services test on iOS
* Make iOS happy by:
* Letting grpcpp_admin conditionally depend on grpcpp_csds
* Fix an unexpected side-effect of dependency update
* generate separate .so with/without debug symbols for grpc_csharp_ext
* add linux .dbginfo.so libraries to Grpc.Core.NativeDebug package
* fix aarch64 grpc_csharp_ext stripping
* lots of linux artifact build jobs are currently timing out
* Migrate the bm_diff benchmarks to python3
Includes a requirements.txt pinned at ~2017 versions, when this script
was first written.
* Replace p2 with p3 scipy/numpy dependencies.
* py2->3 for benchmark setup scripts
* upgrade pip to resolve python3 cryptography/setuptools-rust problem
* re-add jobset import (accidentally removed)
* re-add six's urllib import. This file is still used in py2 tests
* force py3 in run_if_c_cpp_modified.sh
* Fix another instance of subprocess.check_output binary mode
* Use the requirements.txt for CI perf environment setup
* Try to upgrade PyJWT. (v2.0.0 was problematic, #25053)
v2.x makes encode return strs from jwt.encode in both py2 and py3.
Previously, py3 would return bytes, and py2 a str.
* upgate cryptography lib version requirements for jwt.
* Wrap pip requirements specifier in quotes '>=x,<y'
* Decode subprocess output once instead of for every line
* Revert "Decode subprocess output once instead of for every line"
This reverts commit 28d14026431622ac7afd3535a8d7118e5be96628.
py2 doesn't support the `text` argument to subprocess.check_output.
* Address reviewer requests
* Pin a valid scipy version
* Remove scipy and tabulate dependencies from macos tests
Jan Tattermusch
Paulo Castello da Costa
Yash Tibrewal
Wanlin Du
Sergii Tkachenko
Esun Kim
Eric Gribkoff
Richard Belleville
Lidi Zheng
Stanley Cheung
Doug Fawley
AJ Heller