Skip to content

[SPEC]: Test failures on the Linux/ARM64 platform #236

Closed
@odidev

Description

@odidev

Improvement/Addition type

What changes to the current SPEC does this issue focuses upon?

  • [] Add a benchmark variant
  • [] Improve a benchmark variant
  • [] Raise awareness for a new Redis benchmark tool
  • [] Suggest a different Redis Build Variant
  • [ * ] Other (fill details in section bellow)

Detail the requested use-case

I am working with redis-benchmark-specification on the Linux/ARM64 Ubuntu AWS instance.
I am following this documentation < https://github.com/redis/redis-benchmarks-specification#development > to execute tests with tox.
Out of 26 tests, 2 tests are failing, as can be seen from the logs below:

utils/tests/test_self_contained_coordinator.py ....F                                                                                                                                     [ 80%] 
utils/tests/test_self_contained_coordinator_memtier.py F..                                                                                                                               [ 92%] 

These 2 test cases are failing for the same assertion issue (while comparing the value of “result”) after calling the function “test_self_contained_coordinator_blocking_read()” in files test_self_contained_coordinator_memtier.py and test_self_contained_coordinator.py.

=================================================================================== short test summary info ==================================================================================== 
FAILED utils/tests/test_self_contained_coordinator.py::test_self_contained_coordinator_blocking_read - assert False == True 
FAILED utils/tests/test_self_contained_coordinator_memtier.py::test_self_contained_coordinator_blocking_read - assert False == True 
==================================================================== 2 failed, 24 passed, 249 warnings in 567.41s (0:09:27) ==================================================================== 
ERROR: InvocationError for command /home/ubuntu/redis-benchmarks-specification/.tox/integration-tests/bin/coverage run '--include=redis_benchmarks_specification/*' -m pytest -ra (exited with code 1) 

On further debugging, it looks like the tests are failing to run docker container using image redislabs/memtier_benchmark:edge here < https://github.com/redis/redis-benchmarks-specification/blob/main/redis_benchmarks_specification/__self_contained_coordinator__/self_contained_coordinator.py#L1025 >, as this docker image is not available for the Linux/ARM64 platform at DockerHub.

I have added a comment in this already raised ticker < RedisLabs/memtier_benchmark#213 > in the memtier-benchmark GitHub repo, requesting the same.

Can you please provide some pointers whether my analysis seems correct?

Also, may I know, are you interested in adding Linux/ARM64 platform support to redis-benchmark-specification? Kindly provide your suggestions on the same.

Screenshots/CLI snippets of what's expected to happen
If applicable, add screenshots to help explain your problem. For example a chart.

Detail the used benchmark tool

What benchamrk tool to use:

  • [] memtier_benchmark
  • [] redis-benchmark
  • [] redis-benchmark-go
  • [] Other (name it):

Sample benchmark tool command to implement the requested use-case:

$ git clone https://github.com/redis/redis-benchmarks-specification.git && cd redis-benchmarks-specification
$ pip install poetry
$ poetry install
$ poetry run black .
$ poetry run flake8
$ pip3 install -r ./dev_requirements.txt
$ tox

Additional context
Add any other context about the problem here.

Metadata

Metadata

Assignees

Labels

specImprovements or additions to the Redis benchmarks specificationtriageOn triage by Performance Group

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions