Bug #64681
openmgr/dashboard: grpc deps broken in some builds
0%
Description
in fedora, grpcio and grpcio-tools current pinned version is not available. so need to relax on the pinnings.
Updated by Brad Hubbard about 2 months ago
Nizamudeen,
I'm seeing the following failures on Fedora 39 with grpcio and grpcio-tools.
File "/usr/lib64/python3.12/multiprocessing/pool.py", line 48, in mapstar return list(map(*args)) ^^^^^^^^^^^^^^^^ File "/tmp/pip-wheel-9e_3smoc/grpcio_5fdf3d1b28c440f291636dba4227522d/src/python/grpcio/_parallel_compile_patch.py", line 59, in _compile_single_file self._compile(obj, src, ext, cc_args, extra_postargs, pp_opts) File "/home/brad/working/src/cephworking/wip-mon_memory_target-ignored/install-deps-python3/lib64/python3.12/site-packages/setuptools/_distutils/unixccompiler.py", line 187, in _compile raise CompileError(msg) distutils.errors.CompileError: command '/usr/lib64/ccache/gcc' failed with exit code 1 [end of output] note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for grpcio
third_party/protobuf/src/google/protobuf/generated_message_reflection.cc:3342:3: required from here third_party/abseil-cpp/absl/log/internal/check_op.h:336:43: warning: comparison of integer expressions of different signedness: ‘const long unsigned int’ and ‘const int’ [-Wsign-compare] 336 | ABSL_LOG_INTERNAL_CHECK_OP_IMPL(Check_EQ, ==) third_party/abseil-cpp/absl/base/optimization.h:178:58: note: in definition of macro ‘ABSL_PREDICT_TRUE’ 178 | #define ABSL_PREDICT_TRUE(x) (__builtin_expect(false || (x), true)) | ^ third_party/abseil-cpp/absl/log/internal/check_op.h:336:1: note: in expansion of macro ‘ABSL_LOG_INTERNAL_CHECK_OP_IMPL’ 336 | ABSL_LOG_INTERNAL_CHECK_OP_IMPL(Check_EQ, ==) | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ error: command '/usr/lib64/ccache/gcc' failed with exit code 1 [end of output] note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for grpcio-tools
Is this inline with what you are seeing or should I treat this as a separate issue?
In my testing it can be resolved by requiring 1.59.0 at a minimum (1.58.0 and earlier all fail) and I can submit a PR for that if you wish (although maybe we should go to 1.59.3 or even later)?
Updated by Nizamudeen A about 2 months ago
Hi Brad,
Its the same. The reason I didn't pin it to a later version was because in centos8 the latest version available was 1.43.x. And we generated our protobuf files using that version of grpcio-tools. So if we use a later version, it will cause incompatible issues with those protobuf files and break the dashboard nvmeof API. I have created a PR: https://github.com/ceph/ceph/pull/55913/files, which is not working properly so I am checking locally if we can do something differently here.
Updated by Nizamudeen A about 2 months ago
- Status changed from New to In Progress
- Pull request ID set to 55913
Updated by Brad Hubbard about 2 months ago
- Category deleted (
Component - NVMEoF) - Assignee changed from Nizamudeen A to Yehuda Sadeh
- Severity deleted (
3 - minor)
Just FYI, these both fail for me locally as well with updated versions of grpcio and grpcio-tools
run-tox-mgr-dashboard-py3 (Failed)
run-tox-mgr-dashboard-openapi (Failed)