add 'udevadm trigger' wrapper and call after device add rockstor#1606

This was found to be required to ensure lsblk and udev info
was successfully updated after a btrfs device add command.
The failed state of stale info was only rarely observed but
was reproduced a number of times using both whole and
partition pool members.

src/rockstor/storageadmin/views/pool.py

@@ -31,7 +31,7 @@
 from fs.btrfs import (add_pool, pool_usage, resize_pool, umount_root,
                       btrfs_uuid, mount_root, start_balance, usage_bound,
                       remove_share)
-from system.osi import remount
+from system.osi import remount, trigger_udev_update
 from storageadmin.util import handle_exception
 from django.conf import settings
 import rest_framework_custom as rfc
@@ -369,7 +369,8 @@ def put(self, request, pname, command):
                 for d_o in disks:
                     d_o.pool = pool
                     d_o.save()
-
+                # Now we ensure udev info is updated via system wide trigger
+                trigger_udev_update()
             elif (command == 'remove'):
                 if (new_raid != pool.raid):
                     e_msg = ('Raid configuration cannot be changed while '

src/rockstor/system/osi.py

@@ -1691,3 +1691,15 @@ def hostid():
             return puuid
     except:
         return '%s-%s' % (run_command([HOSTID])[0][0], str(uuid.uuid4()))
+
+
+def trigger_udev_update():
+    """
+    In some instances udev info can be out of date after some btrfs opperations.
+    To cause a system wide update of all udev info, and the output of lsblk,
+    we can execute:
+    udevadm trigger
+    This function is a simple wrapper to call the above command via run_command
+    :return: o, e, rc as returned by run_command
+    """
+    return run_command([UDEVADM, 'trigger'])