mgr/cephadm: try again calling ceph-volume without --filter-for-batch

Fixes: https://tracker.ceph.com/issues/48870
This deals with a cephadm upgrade issue:

1. user calls `ceph orch upgrade`
2. mgr/cephadm calls `ceph orch config set mgr.x container_image <new-container>`
3. standby mgr gets upgraded
4. mgr failover to new mgr
5. mgr/cephadm calls `_refresh_host_devices`
6. `_refresh_host_devices` calls` ceph orch config get osd container_image`.
  But this returns the old image
7. `_refresh_host_devices` calls `ceph-volume ... --filter-for-batch`
  with an image that doesn't support `filter-for-batch`

The idea is to simply retiry calling ceph-volume inventory without `--filter-for-batch`

(also removed `out` being used without being declared)

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

diff --git a/src/pybind/mgr/cephadm/serve.py b/src/pybind/mgr/cephadm/serve.py
index f988295fda04aff8c6f75686a51bb6904f679034..75dd4277c763435c4908f45a5260975e48cd269f 100644 (file)
--- a/src/pybind/mgr/cephadm/serve.py
+++ b/src/pybind/mgr/cephadm/serve.py
@@ -293,16 +293,24 @@ class CephadmServe:
 
     def _refresh_host_devices(self, host: str) -> Optional[str]:
         try:
-            devices = self._run_cephadm_json(host, 'osd', 'ceph-volume',
-                                             ['--', 'inventory', '--format=json', '--filter-for-batch'])
+            try:
+                devices = self._run_cephadm_json(host, 'osd', 'ceph-volume',
+                                                 ['--', 'inventory', '--format=json', '--filter-for-batch'])
+            except OrchestratorError as e:
+                if 'unrecognized arguments: --filter-for-batch' in str(e):
+                    devices = self._run_cephadm_json(host, 'osd', 'ceph-volume',
+                                                     ['--', 'inventory', '--format=json'])
+                else:
+                    raise
+
             networks = self._run_cephadm_json(host, 'mon', 'list-networks', [], no_fsid=True)
         except OrchestratorError as e:
             return str(e)
 
         self.log.debug('Refreshed host %s devices (%d) networks (%s)' % (
             host, len(devices), len(networks)))
-        devices = inventory.Devices.from_json(devices)
-        self.mgr.cache.update_host_devices_networks(host, devices.devices, networks)
+        ret = inventory.Devices.from_json(devices)
+        self.mgr.cache.update_host_devices_networks(host, ret.devices, networks)
         self.update_osdspec_previews(host)
         self.mgr.cache.save_host(host)
         return None

diff --git a/src/pybind/mgr/cephadm/tests/test_cephadm.py b/src/pybind/mgr/cephadm/tests/test_cephadm.py
index 840e82d0167e51bb58168204014d5a4464eb314c..12661d413a767a05a4d242a3ce6a631b5f181c04 100644 (file)
--- a/src/pybind/mgr/cephadm/tests/test_cephadm.py
+++ b/src/pybind/mgr/cephadm/tests/test_cephadm.py
@@ -7,6 +7,7 @@ import pytest
 from ceph.deployment.drive_group import DriveGroupSpec, DeviceSelection
 from cephadm.serve import CephadmServe
 from cephadm.services.osd import OSD, OSDRemovalQueue
+from cephadm.utils import CephadmNoImage
 
 try:
     from typing import Any, List
@@ -963,3 +964,31 @@ class TestCephadm(object):
                 assert image == 'image@repo_digest'
             else:
                 assert image == 'image'
+
+    @mock.patch("cephadm.serve.CephadmServe._run_cephadm")
+    def test_ceph_volume_no_filter_for_batch(self, _run_cephadm, cephadm_module: CephadmOrchestrator):
+        _run_cephadm.return_value = ('{}', '', 0)
+
+        error_message = """cephadm exited with an error code: 1, stderr:/usr/bin/podman:stderr usage: ceph-volume inventory [-h] [--format {plain,json,json-pretty}] [path]/usr/bin/podman:stderr ceph-volume inventory: error: unrecognized arguments: --filter-for-batch
+Traceback (most recent call last):
+  File "<stdin>", line 6112, in <module>
+  File "<stdin>", line 1299, in _infer_fsid
+  File "<stdin>", line 1382, in _infer_image
+  File "<stdin>", line 3612, in command_ceph_volume
+  File "<stdin>", line 1061, in call_throws"""
+
+        with with_host(cephadm_module, 'test'):
+            _run_cephadm.reset_mock()
+            _run_cephadm.side_effect = OrchestratorError(error_message)
+
+            s = CephadmServe(cephadm_module)._refresh_host_devices('test')
+            assert s == 'host test `cephadm ceph-volume` failed: ' + error_message
+
+            assert _run_cephadm.mock_calls == [
+                mock.call('test', 'osd', 'ceph-volume',
+                          ['--', 'inventory', '--format=json', '--filter-for-batch'], image='',
+                          no_fsid=False),
+                mock.call('test', 'osd', 'ceph-volume',
+                          ['--', 'inventory', '--format=json'], image='',
+                          no_fsid=False),
+            ]