| 版权：本文版权归作者和博客园共有，欢迎转载，但未经作者同意必须保留此段声明，且在文章页面明显位置给出原文连接。如有问题，可以邮件：wangxu198709@gmail.com
简介：
Multipath：这个多路径软件在Linux平台广泛使用，它的功能就是可以把一个快设备对应的多条路径聚合成一个单一的multipath device。主要目的有如下两点：
多路径冗余（redundancy）：当配置在Active/Passive模式下，只有一半的路径会用来做IO，如果IO路径上有任何失败（包括，交换机故障，线路故障，后端存储故障等），可以自动切换的备用路线上，对上层应用做到基本无感知。
提高性能（Performance）： 当配置在Active/Active模式下，所以路径都可以用来跑IO（如以round-robin模式），可以提高IO速率或者延时。
multipath不是本文的重点，如有需要，请移步：https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/dm_multipath/setup_overview
安装及使用：
Multipath：这个多路径软件在Linux平台广泛使用，在Debian/Ubuntu平台可以通过 sudo apt-get install multipath-tools 安装, RedHat/CentOS 平台可以通过 sudo yum install device-mapper-multipath 安装。
multipath.conf: multipath对于主流的存储阵列都有默认的配置，可以支持存储阵列的很多自带特性，如ALUA。当然用户可以在安装好后，手动创建/etc/multipath.conf
以下是VNX/Unity的参考配置（vnx cinder driver）：
blacklist {
# Skip the files under /dev that are definitely not FC/iSCSI devices
# Different system may need different customization
devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st)[0-9]*"
devnode "^hd[a-z][0-9]*"
devnode "^cciss!c[0-9]d[0-9]*[p[0-9]*]"
# Skip LUNZ device from VNX
device {
vendor "DGC"
product "LUNZ"
}
}
defaults {
user_friendly_names no
flush_on_last_del yes
}
devices {
# Device attributed for EMC CLARiiON and VNX series ALUA
device {
vendor "DGC"
product ".*"
product_blacklist "LUNZ"
path_grouping_policy group_by_prio
path_selector "round-robin 0"
path_checker emc_clariion
features "1 queue_if_no_path"
hardware_handler "1 alua"
prio alua
failback immediate
}
}
Multipath在OpenStack中的应用及faulty device的产生：
OpenStack中，multipath可以使用在Nova和Cinder的节点上，提供对后端存储的高可用访问。在很早之前，这部分代码是分别在Nova和Cinder项目里面的，渐渐的为了维护方便，就单独拧出来一个项目：os-brick
os-brick里面很重要的两个interface是：connect_volume-负责链接一个存储上的LUN或者disk，disconnect_volume-辅助断开与存储上一个LUN的链接。
什么是faulty device
当host上multipath软件发现对应的host path不可访问时，就会显示为faulty状态。
关于所有状态的描述，可以参考：https://en.wikipedia.org/wiki/Linux_DM_Multipath
os-brick的代码我选择的是比较早期容易产生faulty device的版本：https://github.com/openstack/os-brick/blob/liberty-eol/os_brick/initiator/connector.py
1. connect_volume的主要逻辑如下：
@synchronized('connect_volume')
def connect_volume(self, connection_properties):
"""Attach the volume to instance_name.
connection_properties for iSCSI must include:
target_portal(s) - ip and optional port
target_iqn(s) - iSCSI Qualified Name
target_lun(s) - LUN id of the volume
Note that plural keys may be used when use_multipath=True
"""
device_info = {'type': 'block'}
if self.use_multipath:
# Multipath installed, discovering other targets if available
try:
ips_iqns = self._discover_iscsi_portals(connection_properties)
except Exception:
raise exception.TargetPortalNotFound(
target_portal=connection_properties['target_portal'])
if not connection_properties.get('target_iqns'):
# There are two types of iSCSI multipath devices. One which
# shares the same iqn between multiple portals, and the other
# which use different iqns on different portals.
# Try to identify the type by checking the iscsiadm output
# if the iqn is used by multiple portals. If it is, it's
# the former, so use the supplied iqn. Otherwise, it's the
# latter, so try the ip,iqn combinations to find the targets
# which constitutes the multipath device.
main_iqn = connection_properties['target_iqn']
all_portals = set([ip for ip, iqn in ips_iqns])
match_portals = set([ip for ip, iqn in ips_iqns
if iqn == main_iqn])
if len(all_portals) == len(match_portals):
ips_iqns = zip(all_portals, [main_iqn] * len(all_portals))
for ip, iqn in ips_iqns:
props = copy.deepcopy(connection_properties)
props['target_portal'] = ip
props['target_iqn'] = iqn
self._connect_to_iscsi_portal(props)
self._rescan_iscsi()
host_devices = self._get_device_path(connection_properties)
else:
target_props = connection_properties
for props in self._iterate_all_targets(connection_properties):
if self._connect_to_iscsi_portal(props):
target_props = props
break
else:
LOG.warning(_LW(
'Failed to connect to iSCSI portal %(portal)s.'),
{'portal': props['target_portal']})
host_devices = self._get_device_path(target_props)
# The /dev/disk/by-path/... node is not always present immediately
# TODO(justinsb): This retry-with-delay is a pattern, move to utils?
tries = 0
# Loop until at least 1 path becomes available
while all(map(lambda x: not os.path.exists(x), host_devices)):
if tries >= self.device_scan_attempts:
raise exception.VolumeDeviceNotFound(device=host_devices)
LOG.warning(_LW("ISCSI volume not yet found at: %(host_devices)s. "
"Will rescan & retry. Try number: %(tries)s."),
{'host_devices': host_devices,
'tries': tries})
# The rescan isn't documented as being necessary(?), but it helps
72 if self.use_multipath:
73 self._rescan_iscsi()
else:
if (tries):
host_devices = self._get_device_path(target_props)
self._run_iscsiadm(target_props, ("--rescan",))
tries = tries + 1
if all(map(lambda x: not os.path.exists(x), host_devices)):
time.sleep(tries ** 2)
else:
break
if tries != 0:
LOG.debug("Found iSCSI node %(host_devices)s "
"(after %(tries)s rescans)",
{'host_devices': host_devices, 'tries': tries})
# Choose an accessible host device
host_device = next(dev for dev in host_devices if os.path.exists(dev))
93 if self.use_multipath:
94 # We use the multipath device instead of the single path device
95 self._rescan_multipath()
multipath_device = self._get_multipath_device_name(host_device)
if multipath_device is not None:
host_device = multipath_device
LOG.debug("Unable to find multipath device name for "
"volume. Only using path %(device)s "
"for volume.", {'device': host_device})
device_info['path'] = host_device
return device_info
其中重要的逻辑我都用红色标记了，用来发现host上的块设备device
2. disconnect_volume逻辑如下：
@synchronized('connect_volume')
def disconnect_volume(self, connection_properties, device_info):
"""Detach the volume from instance_name.
connection_properties for iSCSI must include:
target_portal(s) - IP and optional port
target_iqn(s) - iSCSI Qualified Name
target_lun(s) - LUN id of the volume
"""
if self.use_multipath:
10 self._rescan_multipath()
host_device = multipath_device = None
host_devices = self._get_device_path(connection_properties)
# Choose an accessible host device
for dev in host_devices:
if os.path.exists(dev):
host_device = dev
multipath_device = self._get_multipath_device_name(dev)
if multipath_device:
break
if not host_device:
LOG.error(_LE("No accessible volume device: %(host_devices)s"),
{'host_devices': host_devices})
raise exception.VolumeDeviceNotFound(device=host_devices)
if multipath_device:
device_realpath = os.path.realpath(host_device)
27 self._linuxscsi.remove_multipath_device(device_realpath)
28 return self._disconnect_volume_multipath_iscsi(
29 connection_properties, multipath_device)
# When multiple portals/iqns/luns are specified, we need to remove
# unused devices created by logging into other LUNs' session.
for props in self._iterate_all_targets(connection_properties):
self._disconnect_volume_iscsi(props)
上面的红色代码块，会把LUN对应的host path从kernel中，和multipath mapper中删除。
3. 竞态Race Condition分析
注意到，以上两个接口都是用的同一个叫（connect_volume)的锁（其实就是用flock实现的Linux上的文件锁）
为了方便描述faulty device的产生，我画了如下的图，来表示两个接口的关系
如上的流程在非并发的情况下是表现正常的，host上的device都可以正常连接和清理。
但是，以上逻辑有个实现上的问题，当高并发情况下，会产生faulty device， 考虑一下执行顺序：
- 右边的disconnect_volume执行完毕，存储上LUN对应的device path(在/dev/disk/by-path下可以看到）和multipath descriptor（multipath -l可以看到）。
- 这个时候，connect_volume锁被释放，左边的connect_volume开始执行，而右边的terminate_connection还没有执行，也就是说，存储上还没有移除host访问LUN的权限，任何host上的scsi rescan还是会发现这个LUN的device。
- 接着，connect_volume按正常执行，iscsi rescan 和multipath rescan都相继执行，造成在步骤 1）中已经删除的device有重新被scan出来。
- 然后，右边的terminate_connection在存储上执行完成，移除了host对LUN的访问，最终就形成的所谓的faulty device，看到的multipath 输出如下(两个multipath descriptor都是faulty的）：
$ sudo multipath -ll 3600601601290380036a00936cf13e711 dm-30 DGC,VRAID
size=2.0G features='1 retain_attached_hw_handler' hwhandler='1 alua' wp=rw
|-+- policy='round-robin 0' prio=0 status=active
| `- 11:0:0:151 sdef 128:112 failed faulty running
`-+- policy='round-robin 0' prio=0 status=enabled
`- 12:0:0:151 sdeg 128:128 failed faulty running 3600601601bd032007c097518e96ae411 dm-2 DGC,VRAID
size=1.0G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
|-+- policy='round-robin 0' prio=0 status=active
`- #:#:#:# - #:# active faulty running
一般来说，有#:#:#:#输出的multipath是可以直接用 sudo multipath -f 3600601601bd032007c097518e96ae411 删除的。
作为第一部分，到这里faulty device的产生介绍完了，后面再找机会，介绍下在os-brick中如何尽量避免faulty device的出现。
参考资料：
RedHat官方multipath的介绍：https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/dm_multipath/mpio_description
EMC VNX driver doc：https://docs.openstack.org/cinder/queens/configuration/block-storage/drivers/dell-emc-vnx-driver.html
Go实现的块设备连接工具：https://github.com/peter-wangxu/goock
iSCSI Faulty Device Cleanup Script for VNX：https://github.com/emc-openstack/vnx-faulty-device-cleanup