first commit

Readme.md

@@ -0,0 +1 @@
+A simple Systemd Daemon writen in python that automatically (force) start your glusterfs volume when all bricks become online again.

pyAutoReupGlusterVolume.py

@@ -0,0 +1,64 @@
+#!/usr/bin/env python3
+
+
+import subprocess
+import os
+import xml.etree.ElementTree as ET
+
+if __name__=="__main__":
+    AllVolumes          = dict()
+    failedVolume        = dict()
+    my_env              = os.environ.copy()    
+    res=subprocess.run("gluster volume status  --xml",shell=True,capture_output=True,universal_newlines=True,env=my_env)
+    root = ET.fromstring(res.stdout)
+    for volume in root.iter('volume'):
+        VolumeName = volume.find('volName').text
+        AllVolumes[VolumeName]=dict()
+        AllVolumes[VolumeName]["nodes"] = dict()
+        for node in volume.findall('node'):
+            NodeName            = node.find('hostname').text
+            Path                = node.find('path').text
+            Port                = node.find('port').text
+            Status              = node.find('status').text
+            pid                 = node.find('pid').text
+            AllVolumes[VolumeName]["nodes"][NodeName] = dict()
+            AllVolumes[VolumeName]["nodes"][NodeName]['Path']     = Path
+            AllVolumes[VolumeName]["nodes"][NodeName]['Port']     = Port
+            AllVolumes[VolumeName]["nodes"][NodeName]['Status']   = Status
+            AllVolumes[VolumeName]["nodes"][NodeName]['pid']      = pid
+            if NodeName == "Self-heal Daemon":
+                continue
+            elif NodeName == "Quota Daemon":
+                continue
+            else:
+                if int(Status)==0 or Port=="N/A":
+                    if not VolumeName in failedVolume:
+                        failedVolume[VolumeName]=[]
+                    failedVolume[VolumeName].append(NodeName)
+        AllVolumes[VolumeName]["tasks"] = dict()
+        for task in volume.findall('task'):
+            type            = task.find('type').text
+            id              = task.find('id').text
+            status          = task.find('status').text
+            statusStr       = task.find('statusStr').text
+            AllVolumes[VolumeName]["tasks"]["type"] = type
+            AllVolumes[VolumeName]["tasks"]["id"] = id
+            AllVolumes[VolumeName]["tasks"]["status"] = status
+            AllVolumes[VolumeName]["tasks"]["statusStr"] = statusStr
+            
+    print("All Volumes:")
+    print(AllVolumes)
+    
+    print("Failed Nodes:")
+    print(failedVolume)
+    
+    for _Volume,_Nodes in failedVolume.items():
+        Pass=True
+        for Node in _Nodes:
+            if subprocess.call(["ping","-q","-c","2",Node]) != 0:
+                Pass=False
+        if Pass==True:
+            print(f" Volume {_Volume} seems online again, force restart")
+            subprocess.call(["gluster","volume","start",_Volume,"force"])
+            
+    print("Finished")

pyAutoReupGlusterVolume.service

@@ -0,0 +1,10 @@
+[Unit]
+Description=Automatic GlusterFs Volume restart
+Wants=pyAutoReupGlusterVolume.timer
+
+[Service]
+Type=oneshot
+ExecStart=/opt/chacha/pyAutoReupGlusterVolume.py
+
+[Install]
+WantedBy=multi-user.target

pyAutoReupGlusterVolume.timer

@@ -0,0 +1,12 @@
+[Unit]
+Description=Automatic GlusterFs Volume restart
+Requires=pyAutoReupGlusterVolume.service
+
+[Timer]
+Unit=pyAutoReupGlusterVolume.service
+OnBootSec=1min
+OnUnitActiveSec=30s
+AccuracySec=1s
+
+[Install]
+WantedBy=timers.target