Repository navigation
Conversation
#17483 added live-retune test cases that pass --router-mode power-of-two to the frontend and a worker. #17899 removed that mode, and the two merged 29 minutes apart, so main's planner unit tests fail on both cases. Use least-loaded for the frontend case and drop the worker case, which the random, least-loaded and --router-mode=kv cases already cover. Refs DYN-5265 Signed-off-by: Karen Chung <karenc@nvidia.com>
|
No actionable comments were generated in the recent review. WalkthroughPlanner tests now expect ChangesRouter mode test expectations
Estimated code review effort: 1 (Trivial) | ~3 minutes Merge Risk:
|
❌ Dynamo PR CI failed — run 41776407217 (attempt 1) on
|
| Other | Jobs |
|---|---|
| dynamo-runtime | ❌ 2 ✅ 5 |
Failure details
2 jobs failed: on both arm64 and amd64, the frontend unit test test_no_registered_env_var_is_read_from_os_environ fails (after 3 auto-retries) because the parsed default for reject_media_status_code is 0 instead of the registered default 415. No cancelled jobs.
❌ dynamo-runtime / test / sequential cuda13.0, arm64: test_no_registered_env_var_is_read_from_os_environ: reject_media_status_code default 0 != 415
Job: dynamo-runtime / test / sequential cuda13.0, arm64 · Failed step: Run CPU-only tests (parallelized) · Logs: gh run view --job 126401376108 -R ai-dynamo/dynamo --log-failed
components/src/dynamo/frontend/tests/test_frontend_parser_builder.py:62: in test_no_registered_env_var_is_read_from_os_environ
assert vars(parser.parse_args([])) == defaults, sentinel
E AssertionError: 0
E Omitting 90 identical items, use -vv to show
E Differing items:
E {'reject_media_status_code': 0} != {'reject_media_status_code': 415}
...
FAILED components/src/dynamo/frontend/tests/test_frontend_parser_builder.py::test_no_registered_env_var_is_read_from_os_environ - AssertionError: 0
== 1 failed, 7784 passed, 170 skipped, 2854 deselected in 1928.40s (0:32:08) ===
components/src/dynamo/frontend/tests/test_frontend_parser_builder.py::test_no_registered_env_var_is_read_from_os_environ is the only failure in the pre_merge_cpu stage (failed on all 3 auto-retries). The frontend argument parser returns reject_media_status_code=0 with no args/env set, while the test expects the registered default of 415; all other 90 parsed defaults match.
❌ dynamo-runtime / test / sequential cuda13.0, amd64: test_no_registered_env_var_is_read_from_os_environ: reject_media_status_code default 0 != 415
Job: dynamo-runtime / test / sequential cuda13.0, amd64 · Failed step: Run CPU-only tests (parallelized) · Logs: gh run view --job 126401376220 -R ai-dynamo/dynamo --log-failed
components/src/dynamo/frontend/tests/test_frontend_parser_builder.py:62: in test_no_registered_env_var_is_read_from_os_environ
assert vars(parser.parse_args([])) == defaults, sentinel
E AssertionError: 0
E Omitting 90 identical items, use -vv to show
E Differing items:
E {'reject_media_status_code': 0} != {'reject_media_status_code': 415}
...
FAILED components/src/dynamo/frontend/tests/test_frontend_parser_builder.py::test_no_registered_env_var_is_read_from_os_environ - AssertionError: 0
== 1 failed, 7784 passed, 170 skipped, 2854 deselected in 1579.91s (0:26:19) ===
components/src/dynamo/frontend/tests/test_frontend_parser_builder.py::test_no_registered_env_var_is_read_from_os_environ is the only failure in the pre_merge_cpu stage (failed on all 3 auto-retries). The frontend argument parser returns reject_media_status_code=0 with no args/env set, while the test expects the registered default of 415; all other 90 parsed defaults match.
For agents
{"pr": 18013, "run_id": 41776407217, "run_attempt": 1, "head_sha": "a8e698f98785a016edc5bc7545d765eef0583ccc", "failures": [{"job": "dynamo-runtime / test / sequential cuda13.0, arm64", "job_id": 126401376108, "failed_step": "Run CPU-only tests (parallelized)", "signature": "AssertionError: reject_media_status_code default 0 != 415", "tests": ["components/src/dynamo/frontend/tests/test_frontend_parser_builder.py::test_no_registered_env_var_is_read_from_os_environ"], "log_cmd": "gh run view --job 126401376108 -R ai-dynamo/dynamo --log-failed"}, {"job": "dynamo-runtime / test / sequential cuda13.0, amd64", "job_id": 126401376220, "failed_step": "Run CPU-only tests (parallelized)", "signature": "AssertionError: reject_media_status_code default 0 != 415", "tests": ["components/src/dynamo/frontend/tests/test_frontend_parser_builder.py::test_no_registered_env_var_is_read_from_os_environ"], "log_cmd": "gh run view --job 126401376220 -R ai-dynamo/dynamo --log-failed"}]}Posted automatically by Devin for run 41776407217. Updated on every full-CI run of this PR.
|
addressed by #14775, closing |
Overview
main's planner unit tests fail on two cases intest_live_retune_frontend_config.pythat pass--router-mode power-of-two, a mode #17899 removed. #17483 added the cases and merged 29 minutes before #17899, after #17899's CI had already run, so each PR passed on its own. This switches the frontend case toleast-loadedand drops the worker case.Key Changes
test_live_retune_frontend_config.py:test_effective_argvpasses--router-mode least-loaded, so one case still passes a hyphenated mode on the command line.WORKER_GOLDEN_CASESdrops--router-mode=power-of-two. Therandom,least-loadedand--router-mode=kvcases still cover a worker that sets its own mode and the--flag=valueform.Validation
test_effective_argvandtest_worker_router_mode_matches_the_real_worker_parserpower-of-twotest_live_retune_frontend_config.pyIssues and Ongoing Work
mainis affected:release/1.7.0still has power-of-two and doesn't have these tests.Pull Request Checklist
/pr-checklistskill.Summary by CodeRabbit
least-loadedrouter mode and worker-parser cases to use the abbreviatedrandomrouter option.