From fb33abaee6e1a4d7e795899819476a3012e0706b Mon Sep 17 00:00:00 2001
From: Andrzej Jackowski <andrzej.jackowski@scylladb.com>
Date: Fri, 15 May 2026 10:16:12 +0200
Subject: [PATCH] test: storage: retry fusermount3 unmount on teardown

After stopping scylla server processes, the FUSE daemon
(fuse2fs) may still be processing file handle closures.
An immediate fusermount3 -u can fail with 'device busy',
causing spurious test failures on teardown.

Retry the unmount up to 10 times with 0.5s delay between
attempts, and capture stderr for diagnostics.

Fixes: SCYLLADB-2066

Closes scylladb/scylladb#29920

(cherry picked from commit 61e5ec98882b3be2cd960b65ed6ef8a2eecad3ae)

Closes scylladb/scylladb#29930
---
 test/cluster/storage/conftest.py | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/test/cluster/storage/conftest.py b/test/cluster/storage/conftest.py
index eb6026475e..8dbe43fda2 100644
--- a/test/cluster/storage/conftest.py
+++ b/test/cluster/storage/conftest.py
@@ -5,10 +5,12 @@
 # SPDX-License-Identifier: LicenseRef-ScyllaDB-Source-Available-1.1
 #
 import pytest
+import logging
 import os
 import pathlib
 import shutil
 import subprocess
+import time
 import uuid
 
 from typing import Callable
@@ -19,6 +21,8 @@ from test.cluster.conftest import PHASE_REPORT_KEY
 from test.pylib.manager_client import ManagerClient
 from test.pylib.util import gather_safely
 
+logger = logging.getLogger(__name__)
+
 
 @pytest.fixture(scope="function")
 def volumes_factory(pytestconfig, build_mode, request):
@@ -64,7 +68,15 @@ def volumes_factory(pytestconfig, build_mode, request):
             shutil.copytree(volume.mount, base.parent.parent / f"scylla-{hash}-{id}", ignore=shutil.ignore_patterns('commitlog*', 'lost+found*'))
             shutil.copyfile(volume.log, base.parent.parent / f"scylla-{hash}-{id}.log")
 
-        subprocess.run(["fusermount3", "-u", volume.mount], check=True)
+        retries = 10
+        for attempt in range(retries):
+            result = subprocess.run(["fusermount3", "-u", volume.mount], capture_output=True, text=True)
+            if result.returncode == 0:
+                break
+            logger.warning("fusermount3 -u attempt %d/%d failed: %s", attempt + 1, retries, result.stderr.strip())
+            if attempt + 1 == retries:
+                raise subprocess.CalledProcessError(result.returncode, result.args, result.stdout, result.stderr)
+            time.sleep(0.5)
         os.unlink(volume.img)