[openshift] Capture ls long listing for /mnt

sos/report/plugins/openshift_lso.py

@@ -0,0 +1,36 @@
+# Copyright (C) 2007-2025 Red Hat, Inc., Jon Magrini <jmagrini@redhat.com>
+
+# This file is part of the sos project: https://github.com/sosreport/sos
+#
+# This copyrighted material is made available to anyone wishing to use,
+# modify, copy, or redistribute it subject to the terms and conditions of
+# version 2 of the GNU General Public License.
+#
+# See the LICENSE file in the source distribution for further information.
+
+from sos.report.plugins import (Plugin, RedHatPlugin)
+
+class OpenshiftLSO(Plugin, RedHatPlugin):
+    """
+    This plugin is used to collect Openshift LSO details. This expands
+    the ceph_osd plugin since storage nodes by default are not setup for
+    ceph access. When gathering data from an OpenShift node when LSO is in use,
+    we currently do not collect the symlink data location for LSO,
+    which is always under /mnt (basically two layers lower,
+    but the names can change). This is useful in determining if LSO
+    is setup correctly when also having a must-gather output.
+    Many times the LSO directory has pointers to paths instead of devices.
+    This can cause issues with lost access to the OCP backend storage
+    used by OpenShift.
+    """
+
+    short_desc = 'Openshift LSO'
+
+    plugin_name = "openshift_lso"
+    profiles = ('storage', 'openshift', 'ceph')
+    # Each node runs the ceph-osd daemon,
+    # which interacts with logical disks attached to the node.
+    files = '/run/ceph/**/ceph-osd*'
+
+    def setup(self):
+        self.add_dir_listing('/mnt', tags=['ls_mnt'], recursive=True)