Merge pull request #50 from xcp-ng/gtn-pr-46

Update smartctl.py

Author: gthvn1

README.md

@@ -152,17 +152,33 @@ $ xe host-call-plugin host-uuid=<uuid> plugin=lsblk.py fn=list_block_devices
 
 ## Smartctl parser
 
-A xapi plugin to get information and health of physical disks on the host
+This XAPI plugin provides information and health details for the physical disks on the host. 
+
+It uses the `smartctl --scan` command to retrieve the list of devices. For devices managed by
+MegaRAID, the device names may be identical. To handle this, the plugin returns information
+for each unique "name:type" pair.
+
+The plugin parses the JSON output from the `smartctl` command to gather information and health
+data. As a result, it requires a version of `smartctl` capable of producing JSON output.
+This functionality is available in **XCP-ng 8.3**, but not in **XCP-ng 8.2**.
+
 ### `information`:
+
+This function returns information about all detected devices. The JSON can be quite big.
+
 ```
 xe host-call-plugin host-uuid=<uuid> plugin=smartctl.py fn=information
-{"/dev/sdf": {"power_on_time": {"hours": 9336}, "ata_version": {"minor_value": 94, "string": "ACS-4 T13/BSR INCITS 529 revision 5", "major_value": 2556}, "form_factor": {"ata_value": 3, "name": "2.5 inches"}, "firmware_version": "SVQ02B6Q", "wwn": {"oui": 9528, "naa": 5, "id": 65536604056}, "smart_status": {"passed": true}, "smartctl": {"build_info": "(local build)", "exit_status": 0, "argv": ["smartctl", "-j", "-a", "/dev/sdf"], "version": [7, 0], "svn_revision": "4883", "platform_info": "x86_64-linux-4.19.0+1"}, "temperature": {"current": 35}, "rotation_rate": 0, "interface_speed": {"current": {"sata_value": 3, "units_per_second": 60, "string": "6.0 Gb/s", "bits_per_unit": 100000000}, [...] }
+{"/dev/nvme1:nvme": {"smart_status": {"nvme": {"value": 0}, "passed": true}, "nvme_controller_id": 0, "smartctl": {"build_info": "(local build)", "exit_status": 0, "argv": ["smartctl", "-j", "-a", "-d", "nvme",
+ "/dev/nvme1"], "version": [7, 0], "svn_revision": "4883", "platform_info": "x86_64-linux-4.19.0+1"}, "temperature": {"current": 32}, ...
 ```
 
 ### `health`:
+
+This function returns health status per detected devices.
+
 ```
 xe host-call-plugin host-uuid=<uuid>  plugin=smartctl.py fn=health
-{"/dev/sdf": "PASSED", "/dev/sdg": "PASSED", "/dev/sdd": "PASSED", "/dev/sde": "PASSED", "/dev/sdb": "PASSED", "/dev/sdc": "PASSED", "/dev/sda": "PASSED"}
+{"/dev/nvme1:nvme": "PASSED", "/dev/sda:scsi": "PASSED", "/dev/nvme0:nvme": "PASSED", "/dev/bus/0:megaraid,1": "PASSED", "/dev/bus/0:megaraid,0": "PASSED"}
 ```
 
 ## Netdata

SOURCES/etc/xapi.d/plugins/smartctl.py

@@ -10,36 +10,37 @@
 from xcpngutils.operationlocker import OperationLocker
 
 @error_wrapped
-def _list_disks():
-    disks = []
+def _list_devices():
+    devices = []
     result = run_command(['smartctl', '--scan'])
     for line in result['stdout'].splitlines():
-        if line.startswith('/dev/') and not line.startswith('/dev/bus/'):
-            disks.append(line.split()[0])
-    return disks
+        devices.append({'name': line.split()[0], 'type': line.split()[2]})
+    return devices
 
 @error_wrapped
 def get_information(session, args):
     results = {}
     with OperationLocker():
-        disks = _list_disks()
-        for disk in disks:
-            cmd = run_command(["smartctl", "-j", "-a", disk], check=False)
-            results[disk] = json.loads(cmd['stdout'])
+        devices = _list_devices()
+        for device in devices:
+            cmd = run_command(["smartctl", "-j", "-a", "-d", device['type'], device['name']], check=False)
+            # We use the name + type as a key because we can have several megaraid with the same name.
+            # So we use the type to differenciate them.
+            results[device['name'] + ":" + device['type']] = json.loads(cmd['stdout'])
         return json.dumps(results)
 
 @error_wrapped
 def get_health(session, args):
     results = {}
     with OperationLocker():
-        disks = _list_disks()
-        for disk in disks:
-            cmd = run_command(["smartctl", "-j", "-H", disk])
+        devices = _list_devices()
+        for device in devices:
+            cmd = run_command(["smartctl", "-j", "-H", "-d", device['type'], device['name']], check=False)
             json_output = json.loads(cmd['stdout'])
             if json_output['smart_status']['passed']:
-                results[disk] = "PASSED"
+                results[device['name'] + ":" + device['type']] = "PASSED"
             else:
-                results[disk] = "FAILED"
+                results[device['name'] + ":" + device['type']] = "FAILED"
         return json.dumps(results)
 
 

tests/smartctl_outputs/__init__.py

@@ -0,0 +1,30 @@
+import json
+from smartctl_outputs.smartctl_sda import INFO_SDA
+from smartctl_outputs.smartctl_nvme1 import INFO_NVME1
+from smartctl_outputs.smartctl_megaraid0 import INFO_MEGARAID0
+from smartctl_outputs.smartctl_megaraid1 import INFO_MEGARAID1
+
+# Parse the INFO JSON string for each devices
+info_sda_dict = json.loads(INFO_SDA)
+info_nvme1_dict = json.loads(INFO_NVME1)
+info_megaraid0_dict = json.loads(INFO_MEGARAID0)
+info_megaraid1_dict = json.loads(INFO_MEGARAID1)
+
+expected_info_dict = {
+    "/dev/sda:sat": info_sda_dict,
+    "/dev/nvme1:nvme": info_nvme1_dict,
+    "/dev/bus/0:megaraid,0": info_megaraid0_dict,
+    "/dev/bus/0:megaraid,1": info_megaraid1_dict,
+}
+
+# Convert the result back to a JSON string
+EXPECTED_INFO = json.dumps(expected_info_dict, indent=2)
+
+expected_health_dict = {
+    "/dev/sda:sat": "PASSED",
+    "/dev/nvme1:nvme": "PASSED",
+    "/dev/bus/0:megaraid,0": "PASSED",
+    "/dev/bus/0:megaraid,1": "PASSED",
+}
+
+EXPECTED_HEALTH = json.dumps(expected_health_dict, indent=2)