diskimage-builder/diskimage_builder/block_device/level3/mount.py

195 lines
7.5 KiB
Python
Raw Permalink Normal View History

# Copyright 2017 Andreas Florath (andreas@florath.net)
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
import functools
import logging
import os
import time
from diskimage_builder.block_device.exception \
import BlockDeviceSetupException
from diskimage_builder.block_device.plugin import NodeBase
from diskimage_builder.block_device.plugin import PluginBase
from diskimage_builder.block_device.utils import exec_sudo
logger = logging.getLogger(__name__)
class MountPointNode(NodeBase):
def __init__(self, mount_base, config, state):
super(MountPointNode, self).__init__(config['name'], state)
# Parameter check
self.mount_base = mount_base
for pname in ['base', 'mount_point']:
if pname not in config:
raise BlockDeviceSetupException(
"MountPoint config needs [%s]" % pname)
setattr(self, pname, config[pname])
logger.debug("MountPoint created [%s]", self)
def get_edges(self):
"""Insert all edges
The dependency edge is created in all cases from the base
element (typically a mkfs) and, if this is not the 'first'
mount-point, an edge is created from the mount-point before in
"sorted order" (see :func:`sort_mount_points`). This ensures
that during mounting (and umounting) the globally correct
order is used.
"""
edge_from = []
edge_to = []
# should have been added by __init__...
assert 'sorted_mount_points' in self.state
sorted_mount_points = self.state['sorted_mount_points']
# If we are not first, add our parent in the global dependency
# list. sorted_mount_points is tuples (mount_point, node_name).
# find ourselves in the mount_points, and our parent node
# is one before us in node_name list.
mount_points = [x[0] for x in sorted_mount_points]
node_name = [x[1] for x in sorted_mount_points]
mpi = mount_points.index(self.mount_point)
if mpi > 0:
dep = node_name[mpi - 1]
edge_from.append(dep)
edge_from.append(self.base)
return (edge_from, edge_to)
def create(self):
if (self.state['filesys'][self.base]['fstype'] == 'swap'):
# No need to mount/activate swap during image build.
mount_point = self.mount_point = 'none'
else:
logger.debug("mount called [%s]", self.mount_point)
rel_mp = self.mount_point if self.mount_point[0] != '/' \
else self.mount_point[1:]
mount_point = os.path.join(self.mount_base, rel_mp)
if not os.path.exists(mount_point):
# Need to sudo this because of permissions in the new
# file system tree.
exec_sudo(['mkdir', '-p', mount_point])
logger.info("Mounting [%s] to [%s]", self.name, mount_point)
exec_sudo(["mount", self.state['filesys'][self.base]['device'],
mount_point])
if 'mount' not in self.state:
self.state['mount'] = {}
self.state['mount'][self.mount_point] \
= {'name': self.name, 'base': self.base, 'path': mount_point}
if 'mount_order' not in self.state:
self.state['mount_order'] = []
self.state['mount_order'].append(self.mount_point)
def umount(self):
if (self.state['filesys'][self.base]['fstype'] == 'swap'):
# Swap not mounted/activated during image build.
return
logger.info("Called for [%s]", self.name)
# Before calling umount, call 'fstrim' on suitable mounted
# file systems. This discards unused blocks from the mounted
# file system and therefore decreases the resulting image
# size.
#
# A race condition can occur when trying to fstrim immediately
# after deleting a file resulting in that free space not being
# reclaimed. Calling sync before fstrim is a workaround for
# this behaviour.
# https://lists.gnu.org/archive/html/qemu-devel/2014-03/msg02978.html
exec_sudo(["sync"])
if self.state['filesys'][self.base]['fstype'] != 'vfat':
exec_sudo(["fstrim", "--verbose",
self.state['mount'][self.mount_point]['path']])
# Even 'fstrim' call sometimes don't solve the issue with 'busy'
# filesystem, so we need to catch the exception and repeat unount.
mount_point = self.state['mount'][self.mount_point]['path']
catch = None
for try_cnt in range(10, 1, -1):
try:
exec_sudo(["umount", mount_point])
return
except BlockDeviceSetupException as e:
catch = e
logger.error("umount failed (%s)", e.output.strip())
time.sleep(3)
logger.debug("Gave up trying to umount [%s]", mount_point)
raise catch
def delete(self):
self.umount()
def cmp_mount_order(this, other):
"""Sort comparision function for mount-point sorting
See if ``this`` comes before ``other`` in mount-order list. In
words: if the other mount-point has us as it's parent, we come
before it (are less than it). e.g. ``/var < /var/log <
/var/log/foo``
:param this: tuple of mount_point, node name
:param other: tuple of mount_point, node name
:returns int: cmp value
"""
# sort is only based on the mount_point.
this, _ = this
other, _ = other
if this == other:
return 0
if other.startswith(this):
return -1
else:
return 1
class Mount(PluginBase):
def __init__(self, config, defaults, state):
super(Mount, self).__init__()
if 'mount-base' not in defaults:
raise BlockDeviceSetupException(
"Mount default config needs 'mount-base'")
self.node = MountPointNode(defaults['mount-base'], config, state)
# save this new node to the global mount-point list and
# re-order it to keep it in mount-order. Used in get_edges()
# to ensure we build the mount graph in order
#
# note we can't just put the MountPointNode into the state,
# because it's not json serialisable and we still dump the
# state to json. that's why we have this (mount_point, name)
# tuples and sorting trickery
sorted_mount_points = state.get('sorted_mount_points', [])
mount_points = [mp for mp, name in sorted_mount_points]
if self.node.mount_point in mount_points:
raise BlockDeviceSetupException(
"Mount point [%s] specified more than once"
% self.node.mount_point)
sorted_mount_points.append((self.node.mount_point, self.node.name))
sorted_mount_points.sort(key=functools.cmp_to_key(cmp_mount_order))
# Save the state if it's new (otherwise this is idempotent update)
state['sorted_mount_points'] = sorted_mount_points
logger.debug("Ordered mounts now: %s", sorted_mount_points)
def get_nodes(self):
return [self.node]