Fix validation workflow on aarch64 with conda 23.11.0 and GLIBC_2.25 (#…

…1645) * Debug aarch64 clone * Debug * Fix validation workflow with conda 23.11.0 and GLIBC_2.25 * Gate the change on linux-aarch64 and keep the old LD_LIBRARY_PATH * Try to unset LD_LIBRARY_PATH in the workflow instead * Fix copy/paste typo
pytorch · Dec 16, 2023 · cd257e9 · cd257e9
1 parent df2d4e8
commit cd257e9
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 2 deletions.
diff --git a/.github/workflows/validate-aarch64-linux-binaries.yml b/.github/workflows/validate-aarch64-linux-binaries.yml
@@ -95,5 +95,11 @@ jobs:
         printf '%s\n' ${{ toJson(inputs.release-matrix) }} > release_matrix.json
         eval "$(conda shell.bash hook)"
 
-        # Standart case: Validate binaries
+        # NB: The latest conda 23.11.0 pulls in some dependencies of conda-libmamba-solver that
+        # require GLIBC_2.25, which is not available in the current aarch64 image causing the
+        # subsequence git command to fail. Basically, they don't work with CentOS 7 which AML 2
+        # is based on https://github.com/ContinuumIO/anaconda-issues/issues/12822
+        unset LD_LIBRARY_PATH
+
+        # Standard case: Validate binaries
         source ./.github/scripts/validate_binaries.sh
diff --git a/test/smoke_test/smoke_test.py b/test/smoke_test/smoke_test.py
@@ -258,7 +258,16 @@ def smoke_test_modules():
         if module["repo"]:
             if not os.path.exists(f"{cwd}/{module['repo_name']}"):
                 print(f"Path does not exist: {cwd}/{module['repo_name']}")
-                subprocess.check_output(f"git clone --depth 1 {module['repo']}", stderr=subprocess.STDOUT, shell=True)
+                try:
+                    subprocess.check_output(
+                        f"git clone --depth 1 {module['repo']}",
+                        stderr=subprocess.STDOUT,
+                        shell=True,
+                    )
+                except subprocess.CalledProcessError as exc:
+                    raise RuntimeError(
+                        f"Cloning {module['repo']} FAIL: {exc.returncode} Output: {exc.output}"
+                    ) from exc
             try:
                 smoke_test_command = f"python3 {module['smoke_test']}"
                 if target_os == 'windows':