Menu Close
Settings Close

Language and Page Formatting Options

Red Hat Training

A Red Hat training course is available for RHEL 8

Configuring device mapper multipath

Red Hat Enterprise Linux 8

Using the Device Mapper Multipath feature

Red Hat Customer Content Services

Abstract

This documentation collection provides instructions on how to configure and manage the Device Mapper Multipath (DM-Multipath) feature on Red Hat Enterprise Linux 8.

Making open source more inclusive

Red Hat is committed to replacing problematic language in our code, documentation, and web properties. We are beginning with these four terms: master, slave, blacklist, and whitelist. Because of the enormity of this endeavor, these changes will be implemented gradually over several upcoming releases. For more details, see our CTO Chris Wright’s message.

Providing feedback on Red Hat documentation

We appreciate your feedback on our documentation. Let us know how we can improve it.

Submitting comments on specific passages

  1. View the documentation in the Multi-page HTML format and ensure that you see the Feedback button in the upper right corner after the page fully loads.
  2. Use your cursor to highlight the part of the text that you want to comment on.
  3. Click the Add Feedback button that appears near the highlighted text.
  4. Add your feedback and click Submit.

Submitting feedback through Bugzilla (account required)

  1. Log in to the Bugzilla website.
  2. Select the correct version from the Version menu.
  3. Enter a descriptive title in the Summary field.
  4. Enter your suggestion for improvement in the Description field. Include links to the relevant parts of the documentation.
  5. Click Submit Bug.

Chapter 1. Overview of device mapper multipathing

With Device mapper multipathing (DM Multipath), you can configure multiple I/O paths between server nodes and storage arrays into a single device. These I/O paths are physical Storage Area Network (SAN) connections that can include separate cables, switches, and controllers. Multipathing aggregates the I/O paths and creates a new device that consists of the aggregated paths.

DM Multipath provides:

Redundancy
DM Multipath can provide failover in an active/passive configuration. In an active/passive configuration, only half the paths are used at any time for I/O. If any element of an I/O path such as the cable, switch, or controller fails, DM Multipath switches to an alternate path.
Improved Performance
DM Multipath can be configured in an active/active mode, where I/O is spread over the paths in a round-robin fashion. In some configurations, DM Multipath can detect loading on the I/O paths and dynamically rebalance the load.

1.1. Active/Passive multipath configuration with one RAID device

In this configuration, there are two Host Bus Adapters (HBAs) on the server, two SAN switches, and two RAID controllers. Following are the possible failure in this configuration:

  • HBA failure
  • Fibre Channel cable failure
  • SAN switch failure
  • Array controller port failure

With DM Multipath configured, a failure at any of these points causes DM Multipath to switch to the alternate I/O path. The following image describes the configuration with two I/O paths from the server to a RAID device. Here, there is one I/O path that goes through hba1, SAN1, and cntrlr1 and a second I/O path that goes through hba2, SAN2, and cntrlr2.

Figure 1.1. Active/Passive multipath configuration with one RAID device

Active/Passive multipath configuration with one RAID device

1.2. Active/Passive multipath configuration with two RAID devices

In this configuration, there are two HBAs on the server, two SAN switches, and two RAID devices with two RAID controllers each. With DM Multipath configured, a failure at any of the points of the I/O path to either of the RAID devices causes DM Multipath to switch to the alternate I/O path for that device. The following image describes the configuration with two I/O paths to each RAID device. Here, there are two I/O paths to each RAID device.

Figure 1.2. Active/Passive multipath configuration with two RAID device

Active/Passive multipath configuration with two RAID device

1.3. Active/Active multipath configuration with one RAID device

In this configuration, there are two HBAs on the server, two SAN switches, and two RAID controllers. The following image describes the configuration with two I/O paths from the server to a storage device. Here, I/O can be spread among these two paths.

Figure 1.3. Active/Active multipath configuration with one RAID device

Active/Active multipath configuration with one RAID device

1.4. DM Multipath components

The following table describes the DM Multipath components.

Table 1.1. Components of DM Multipath

Component

Description

dm_multipath kernel module

Reroutes I/O and supports failover for paths and path groups.

mpathconf utility

Configures and enables device mapper multipathing.

multipath command

Lists and configures the multipath devices. It is also executed by udev whenever a block device is added, to determine if the device should be part of a multipath device or not.

multipathd daemon

Automatically creates and removes multipath devices and monitors paths; as paths fail and come back, it may update the multipath device. Allows interactive changes to multipath devices. Reload the service if there are any changes to the /etc/multipath.conf file.

kpartx command

Creates device mapper devices for the partitions on a device. This command is automatically executed by udev when multipath devices are created to create partition devices on top of them. The kpartx command is provided in its own package, but the device-mapper-multipath package depends on it.

mpathpersist

Sets up SCSI-3 persistent reservations on multipath devices. This command works similarly to the way sg_persist works for SCSI devices that are not multipathed, but it handles setting persistent reservations on all paths of a multipath device. It coordinates with multipathd to ensure that the reservations are set up correctly on paths that are added later. To use this functionality, the reservation_key attribute must be defined in the /etc/multipath.conf file. Otherwise the multipathd daemon will not check for persistent reservations for newly discovered paths or reinstated paths.

1.5. The multipath command

The multipath command is used to detect and combine multiple paths to devices. It provides a variety of options you can use to administer your multipathed devices.

The following table describes some options of the multipath command that you may find useful.

Table 1.2. Useful multipath command options

OptionDescription

-l

Display the current multipath configuration gathered from sysfs and the device mapper.

-ll

Display the current multipath configuration gathered from sysfs, the device mapper, and all other available components on the system.

-f device

Remove the named multipath device.

-F

Remove all unused multipath devices.

-w device

Remove the wwid of the specified device from the wwids file.

-W

Reset the wwids file to include only the current multipath devices.

1.6. Multipath command output

When you create, modify, or list a multipath device, you get a display of the current device setup. The format is as follows.

  • For each multipath device:
action_if_any: alias (wwid_if_different_from_alias) dm_device_name_if_known vendor,product size=size features='features' hwhandler='hardware_handler' wp=write_permission_if_known
  • For each path group:
-+- policy='scheduling_policy' prio=prio_if_known status=path_group_status_if_known
  • For each path:
 `- host:channel:id:lun devnode major:minor dm_status_if_known path_status online_status

For example, the output of a multipath command might appear as follows:

3600d0230000000000e13955cc3757800 dm-1 WINSYS,SF2372
size=269G features='0' hwhandler='0' wp=rw
|-+- policy='round-robin 0' prio=1 status=active
| `- 6:0:0:0 sdb 8:16  active ready  running
`-+- policy='round-robin 0' prio=1 status=enabled
  `- 7:0:0:0 sdf 8:80  active ready  running

If the path is up and ready for I/O, the status of the path is ready or ghost. If the path is down, the status is faulty or shaky. The path status is updated periodically by the multipathd daemon based on the polling interval defined in the /etc/multipath.conf file.

Additional possible path status values are as follows.

  • i/o pending: The checker is actively checking this path, and the state will be updated shortly.
  • i/o timeout: This is the same as faulty. It lets the user know that the checker did not return either success or failure before the timeout period.
  • removed: The path has been removed from the system, and will shortly be removed from the multipath device. It is treated the same as faulty.
  • wild: multipathd was unable to run the path checker, because of an internal error or configuration issue. This is roughly the same as faulty, except multipath will skip many actions on the path.
  • unchecked: The path checker has not run on this path, either because it has just been discovered, it does not have an assigned path checker, or the path checker encountered an error. This is treated the same as wild.
  • delayed: The path checker returns that the path is up, but multipath is delaying the reinstatement of the path because the path has recently failed multiple times and multipath has been configured to delay paths in this case.

In terms of the kernel, the dm status is similar to the path status The active dm state covers the ready and ghost path states. The pending path state has no equivalent dm state. All other path states map to the failed dm state. The dm state will retain its current status until the path checker has completed.

The possible values for online_status are running and offline. The offline status means that this SCSI device has been disabled.

Note

When you create or modify a multipath device, multipath prints the device configuration. However, some of the features, for example, write permissions, and other feature information might be unknown. There might be a difference between the output and the features that you selected during creation or modification. This is normal behaviour. List the device after creation to view the correct state.

1.7. Displaying multipath configuration

You can use the -l and multipath command to display the current multipath configuration. The -l option displays multipath topology gathered from information in sysfs and the device mapper. The -ll option displays the information the -l option displays in addition to all other available components of the system.

When displaying the multipath configuration, you can specify a verbosity level with the -v option of the multipath command. Specifying -v0 yields no output. Specifying -v1 outputs the created or updated multipath names only, which you can then feed to other tools such as kpartx. Specifying -v2 prints all detected paths, multipaths, and device maps. For even more detailed information, you can also specify -v3, -v4, or -v5.

The following example shows the output of a multipath -l command.

# multipath -l
3600d0230000000000e13955cc3757800 dm-1 WINSYS,SF2372
size=269G features='0' hwhandler='0' wp=rw
|-- policy='round-robin 0' prio=1 status=active
| `- 6:0:0:0 sdb 8:16  active ready  running
`-- policy='round-robin 0' prio=1 status=enabled
  `- 7:0:0:0 sdf 8:80  active ready  running

The following example shows the output of a multipath -ll command.

# multipath -ll
3600d0230000000000e13955cc3757801 dm-10 WINSYS,SF2372
size=269G features='0' hwhandler='0' wp=rw
|-- policy='round-robin 0' prio=1 status=enabled
| `- 19:0:0:1 sdc 8:32  active ready  running
`-- policy='round-robin 0' prio=1 status=enabled
  `- 18:0:0:1 sdh 8:112 active ready  running
3600d0230000000000e13955cc3757803 dm-2 WINSYS,SF2372
size=125G features='0' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=1 status=active
  |- 19:0:0:3 sde 8:64  active ready  running
  `- 18:0:0:3 sdj 8:144 active ready  running

1.8. Additional resources

  • multipath(8) and multipathd(8) man pages
  • /etc/multipath.conf file

Chapter 2. Multipath devices

DM Multipath provides a way of organizing the I/O paths logically, by creating a single multipath device on top of the underlying devices. Without DM Multipath, system treats each path from a server node to a storage controller as a separate device, even when the I/O path connects the same server node to the same storage controller.

2.1. Multipath device identifiers

When new devices are under the control of DM Multipath, these devices are created in the /dev/mapper/ and /dev/ directory.

Note

Any devices of the form /dev/dm-X are for internal use only and should never be used by the administrator directly.

The following describes multipath device names:

  • When the user_friendly_names configuration option is set to no, the name of the multipath device is set to World Wide Identifier (WWID). By default, the name of a multipath device is set to its WWID. The device name would be /dev/mapper/WWID. It is also created in the /dev/ directory, named as /dev/dm-X.
  • Alternately, you can set the user_friendly_names option to yes in the /etc/multipath.conf file. This sets the alias in the multipath section to a node-unique name of the form mpathN. The device name would be /dev/mapper/mpathN and /dev/dm-X. But the device name is not guaranteed to be the same on all nodes using the multipath device. Similarly, if you set the alias option in the /etc/multipath.conf file, the name is not automatically consistent across all nodes in the cluster.
Note

This should not cause any difficulties if you use LVM to create logical devices from the multipath device. To keep your multipath device names consistent in every node, Red Hat recommends disabling the user_friendly_names option.

For example, a node with two HBAs attached to a storage controller with two ports by means of a single unzoned FC switch sees four devices: /dev/sda, /dev/sdb, /dev/sdc, and /dev/sdd. DM Multipath creates a single device with a unique WWID that reroutes I/O to those four underlying devices according to the multipath configuration.

In addition to the user_friendly_names and alias options, a multipath device also has other attributes. You can modify these attributes for a specific multipath device by creating an entry for that device in the multipaths section of the /etc/multipath.conf file.

Additional resources

2.2. Multipath devices in logical volumes

After creating multipath devices, you can use the multipath device names as you would use a physical device name when creating an Logical volume manager (LVM) physical volume. For example, if /dev/mapper/mpatha is the name of a multipath device, the pvcreate /dev/mapper/mpatha command marks /dev/mapper/mpatha as a physical volume.

You can use the resulting LVM physical device when you create an LVM volume group just as you would use any other LVM physical device.

To filter all the sd devices in the /etc/lvm/lvm.conf file, add the filter = [ "r/block/", "r/disk/", "r/sd./", "a/./" ] filter in the devices section of the file.

Note

If you attempt to create an LVM physical volume on a whole device on which you have configured partitions, the pvcreate command fails. The Anaconda and Kickstart installation programs create empty partition tables if you do not specify otherwise for every block device. If you want to use the whole device instead of creating a partition, remove the existing partitions from the device. You can remove existing partitions with the kpartx -d device command and the fdisk utility. If your system has block devices that are greater than 2Tb, use the parted utility to remove partitions.

When you create an LVM logical volume that uses active/passive multipath arrays as the underlying physical devices, you can optionally include filters in the /etc/lvm/lvm.conf file to exclude the disks that underline the multipath devices. This is because if the array automatically changes the active path to the passive path when it receives I/O, multipath will failover and failback whenever LVM scans the passive path, if these devices are not filtered.

The kernel changes the active/passive state by automatically detecting the correct hardware handler to use. For active/passive paths that require intervention to change their state, multipath automatically uses this hardware handler to do so as necessary. If the kernel does not automatically detect the correct hardware handler to use, you can configure which hardware handler to use in the multipath.conf file with the "hardware_handler" option. For active/passive arrays that require a command to make the passive path active, LVM prints a warning message when this occurs.

Depending on your configuration, LVM may print any of the following messages:

  • LUN not ready:

    end_request: I/O error, dev sdc, sector 0
    sd 0:0:0:3: Device not ready: <6>: Current: sense key: Not Ready
        Add. Sense: Logical unit not ready, manual intervention required
  • Read failed:

    /dev/sde: read failed after 0 of 4096 at 0: Input/output error

The following are the reasons for the mentioned errors:

  • Multipath is not set up on storage devices that are providing active/passive paths to a machine.
  • Paths are accessed directly, instead of through the multipath device.

Additional resources

Chapter 3. Configuring DM Multipath

You can set up DM Multipath with the mpathconf utility. This utility creates or edits the /etc/multipath.conf multipath configuration file based on the following scenarios:

  • If the /etc/multipath.conf file already exists, the mpathconf utility will edit it.
  • If the /etc/multipath.conf file does not exist, the mpathconf utility will create the /etc/multipath.conf file from scratch.

3.1. Checking for the device-mapper-multipath package

Before setting up DM Multipath on your system, ensure that your system is up-to-date and includes the device-mapper-multipath package.

Procedure

  1. Check if your system includes the device-mapper-multipath package:

    # rpm -q device-mapper-multipath
    device-mapper-multipath-current-package-version

    If your system does not include the package, it prints the following:

    package device-mapper-multipath is not installed
  2. If your system does not include the package, install it by running the following command:

    # {PackageManager} install device-mapper-multipath

3.2. Setting up DM Multipath for a basic failover configuration

Use the following procedure to set up DM Multipath for a basic failover configuration if you need to edit the /etc/multipath.conf file before starting the multipathd daemon.

Procedure

  1. Enable the multipath configuration file:

    # mpathconf --enable
  2. Edit the /etc/multipath.conf file if necessary. The default settings for DM Multipath are compiled into the system and do not need to be explicitly set in the /etc/multipath.conf file.

    The default value of path_grouping_policy is set to failover, so in this example you do not need to edit the /etc/multipath.conf file.

    The initial defaults section of the configuration file configures your system so that the names of the multipath devices are of the form /dev/mapper/mpathn; without this setting, the names of the multipath devices would be aliased to the WWID of the device. If you do not want to use user friendly names, you can enter the following command:

    # mpathconf --enable --user_friendly_names n

    If you need to edit the multipath configuration file after you have started the multipath daemon, you must execute the systemctl reload multipathd.service command for the changes to take effect.

  3. Save the configuration file and exit the editor, if necessary.
  4. Start the multipath daemon and create the multipath devices:

    # systemctl start multipathd.service
Note

If you remove the device-mapper-multipath package, this does not remove the /etc/multipath.conf file, or any files in the /etc/multipath directory, since that directory can contain more that just the files listed currently. You might need to remove those files manually on subsequent installations of the device-mapper-multipath package.

3.3. Ignoring local disks when generating multipath devices

Some machines have local SCSI cards for their internal disks and DM Multipath is not recommended for these devices. If you set the find_multipaths configuration parameter to on, you do not have to disable multipathing on these devices.

If you do not set the find_multipaths configuration parameter to on, you can use the following procedure to modify the DM Multipath configuration file to ignore the local disks when configuring multipath.

Procedure

  1. Determine which disks are the internal disks. In these examples, /dev/sda is the internal disk:

    • Display existing multipath devices:

      # multipath -v2 -l
      
      SIBM-ESXSST336732LC____F3ET0EP0Q000072428BX1 dm-2 WINSYS,SF2372
      size=33 GB features="0" hwhandler="0" wp=rw
      `-+- policy='round-robin 0' prio=0 status=active
        |- 0:0:0:0 sda 8:0 active undef running
    • Display additional multipath devices that DM Multipath could create:

      # multipath -v2 -d
      
      : SIBM-ESXSST336732LC____F3ET0EP0Q000072428BX1 undef WINSYS,SF2372
      size=33 GB features="0" hwhandler="0" wp=undef
      `-+- policy='round-robin 0' prio=1 status=undef
        |- 0:0:0:0 sda 8:0  undef ready running
  2. Edit the blacklist section of the /etc/multipath.conf file to include this device.

    Identify the device using its WWID attribute. Although you could identify the sda device using a devnode type, that would not be a safe procedure because /dev/sda is not guaranteed to be the same on reboot.

    In the previous example, the WWID of the /dev/sda device is SIBM-ESXSST336732LC____F3ET0EP0Q000072428BX1. To ignore this device, include the following in the /etc/multipath.conf file:

    blacklist {
          wwid SIBM-ESXSST336732LC____F3ET0EP0Q000072428BX1
    }
  3. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  4. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

Additional resources

  • multipath.conf(5) man page

3.4. Configuring additional storage devices

By default, DM Multipath includes support for the most common storage arrays, which support DM Multipath.

Procedure

  • View the default configuration value, including supported devices:

    # multipathd show config
    # multipath -t
  • Optional: To add an additional storage device that is not supported by default as a known multipath device, edit the /etc/multipath.conf file and insert the appropriate device information.

    The following example ilustrates how to add information about the HP Open-V series. This sets the device to queue for a minute or 12 retries and 5 seconds per retry after all paths have failed.

    devices {
            device {
                    vendor "HP"
                    product "OPEN-V"
                    no_path_retry 12
            }
    }

3.5. Setting up multipathing in the initramfs file system

You can set up multipathing in the initramfs file system. You do not need to set it up if you are not using the devices that you want multipathed, until after boot leaves the initramfs file system.

Prerequisites

  • You have configured DM multipath in your system.

Procedure

  • Rebuild the initramfs file system with the multipath configuration files by running the following command:

    # dracut --force --add multipath

    If you run multipath from the initramfs file system and you make any changes to the multipath configuration files, you must rebuild the initramfs file system for the changes to take effect. When a root device uses multipath, running the dracut command automatically adds the multipath module to the initramfs.

  • Optional: If you need multipath running in the initramfs, but you are not setting up a multipathed root device, run:

    # echo add_dracutmodules+=\"multipath\" > /etc/dracut.conf.d/multipath.conf
    # dracut --force
Note

The dracut command includes multipath in the initramfs, even after multipath is no longer required. To stop including multipath, run:

# rm /etc/dracut.conf.d/multipath.conf
# dracut --force

Chapter 4. Enabling multipathing on NVMe devices

You can multipath NVMe devices that are connected to your system over a fabric transport, such as Fibre Channel (FC). You can select between multiple multipathing solutions.

4.1. Native NVMe multipathing and DM Multipath

NVMe devices support a native multipathing functionality. When configuring multipathing on NVMe, you can select between the standard DM Multipath framework and the native NVMe multipathing.

Both DM Multipath and native NVMe multipathing support the Asymmetric Namespace Access (ANA) multipathing scheme of NVMe devices. ANA identifies optimized paths between the controller and the host, and improves performance.

When native NVMe multipathing is enabled, it applies globally to all NVMe devices. It can provide higher performance, but does not contain all of the functionality that DM Multipath provides. For example, native NVMe multipathing supports only the numa and round-robin path selection methods.

Red Hat recommends that you use DM Multipath in Red Hat Enterprise Linux 8 as your default multipathing solution.

4.2. Enabling native NVMe multipathing

This procedure enables multipathing on connected NVMe devices using the native NVMe multipathing solution.

Prerequisites

Procedure

  1. Check if native NVMe multipathing is enabled in the kernel:

    # cat /sys/module/nvme_core/parameters/multipath

    The command displays one of the following:

    N
    Native NVMe multipathing is disabled.
    Y
    Native NVMe multipathing is enabled.
  2. If native NVMe multipathing is disabled, enable it using one of the following methods:

    • Using a kernel option:

      1. Add the nvme_core.multipath=Y option on the kernel command line:

        # grubby --update-kernel=ALL --args="nvme_core.multipath=Y"
      2. On the 64-bit IBM Z architecture, update the boot menu:

        # zipl
      3. Reboot the system.
    • Using a kernel module configuration file:

      1. Create the /etc/modprobe.d/nvme_core.conf configuration file with the following content:

        options nvme_core multipath=Y
      2. Back up the initramfs file system:

        # cp /boot/initramfs-$(uname -r).img \
             /boot/initramfs-$(uname -r).bak.$(date +%m-%d-%H%M%S).img
      3. Rebuild the initramfs file system:

        # dracut --force --verbose
      4. Reboot the system.
  3. Optional: On the running system, change the I/O policy on NVMe devices to distribute the I/O on all available paths:

    # echo "round-robin" > /sys/class/nvme-subsystem/nvme-subsys0/iopolicy
  4. Optional: Set the I/O policy persistently using udev rules. Create the /etc/udev/rules.d/71-nvme-io-policy.rules file with the following content:

    ACTION=="add|change", SUBSYSTEM=="nvme-subsystem", ATTR{iopolicy}="round-robin"

Verification

  1. Check that your system recognizes the NVMe devices:

    # nvme list
    
    Node             SN                   Model                                    Namespace Usage                      Format           FW Rev
    ---------------- -------------------- ---------------------------------------- --------- -------------------------- ---------------- --------
    /dev/nvme0n1     a34c4f3a0d6f5cec     Linux                                    1         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
    /dev/nvme0n2     a34c4f3a0d6f5cec     Linux                                    2         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
  2. List all connected NVMe subsystems:

    # nvme list-subsys
    
    nvme-subsys0 - NQN=testnqn
    \
     +- nvme0 fc traddr=nn-0x20000090fadd597a:pn-0x10000090fadd597a host_traddr=nn-0x20000090fac7e1dd:pn-0x10000090fac7e1dd live
     +- nvme1 fc traddr=nn-0x20000090fadd5979:pn-0x10000090fadd5979 host_traddr=nn-0x20000090fac7e1dd:pn-0x10000090fac7e1dd live
     +- nvme2 fc traddr=nn-0x20000090fadd5979:pn-0x10000090fadd5979 host_traddr=nn-0x20000090fac7e1de:pn-0x10000090fac7e1de live
     +- nvme3 fc traddr=nn-0x20000090fadd597a:pn-0x10000090fadd597a host_traddr=nn-0x20000090fac7e1de:pn-0x10000090fac7e1de live

    Check the active transport type. For example, nvme0 fc indicates that the device is connected over the Fibre Channel transport, and nvme tcp indicates that the device is connected over TCP.

  3. If you edited the kernel options, check that native NVMe multipathing is enabled on the kernel command line:

    # cat /proc/cmdline
    
    BOOT_IMAGE=[...] nvme_core.multipath=Y
  4. Check that DM Multipath reports the NVMe namespaces as, for example, nvme0c0n1 through nvme0c3n1, and not as, for example, nvme0n1 through nvme3n1:

    # multipath -e -ll | grep -i nvme
    
    uuid.8ef20f70-f7d3-4f67-8d84-1bb16b2bfe03 [nvme]:nvme0n1 NVMe,Linux,4.18.0-2
    | `- 0:0:1    nvme0c0n1 0:0     n/a   optimized live
    | `- 0:1:1    nvme0c1n1 0:0     n/a   optimized live
    | `- 0:2:1    nvme0c2n1 0:0     n/a   optimized live
      `- 0:3:1    nvme0c3n1 0:0     n/a   optimized live
    
    uuid.44c782b4-4e72-4d9e-bc39-c7be0a409f22 [nvme]:nvme0n2 NVMe,Linux,4.18.0-2
    | `- 0:0:1    nvme0c0n1 0:0     n/a   optimized live
    | `- 0:1:1    nvme0c1n1 0:0     n/a   optimized live
    | `- 0:2:1    nvme0c2n1 0:0     n/a   optimized live
      `- 0:3:1    nvme0c3n1 0:0     n/a   optimized live
  5. If you changed the I/O policy, check that round-robin is the active I/O policy on NVMe devices:

    # cat /sys/class/nvme-subsystem/nvme-subsys0/iopolicy
    
    round-robin

4.3. Enabling DM Multipath on NVMe devices

This procedure enables multipathing on connected NVMe devices using the DM Multipath solution.

Prerequisites

Procedure

  1. Check that native NVMe multipathing is disabled:

    # cat /sys/module/nvme_core/parameters/multipath

    The command displays one of the following:

    N
    Native NVMe multipathing is disabled.
    Y
    Native NVMe multipathing is enabled.
  2. If native NVMe multipathing is enabled, disable it:

    1. Remove the nvme_core.multipath=Y option from the kernel command line:

      # grubby --update-kernel=ALL --remove-args="nvme_core.multipath=Y"
    2. On the 64-bit IBM Z architecture, update the boot menu:

      # zipl
    3. Remove the options nvme_core multipath=Y line from the /etc/modprobe.d/nvme_core.conf file, if it is present.
    4. Reboot the system.
  3. Make sure that DM Multipath is enabled:

    # systemctl enable --now multipathd.service
  4. Distribute I/O on all available paths. Add the following content in the /etc/multipath.conf file:

    device {
      vendor "NVME"
      product ".*"
      path_grouping_policy    group_by_prio
    }
    Note

    The /sys/class/nvme-subsystem/nvme-subsys0/iopolicy configuration file has no effect on the I/O distribution when DM Multipath manages the NVMe devices.

  5. Reload the multipathd service to apply the configuration changes:

    # multipath -r
  6. Back up the initramfs file system:

    # cp /boot/initramfs-$(uname -r).img \
         /boot/initramfs-$(uname -r).bak.$(date +%m-%d-%H%M%S).img
  7. Rebuild the initramfs file system:

    # dracut --force --verbose

Verification

  1. Check that your system recognizes the NVMe devices:

    # nvme list
    
    Node             SN                   Model                                    Namespace Usage                      Format           FW Rev
    ---------------- -------------------- ---------------------------------------- --------- -------------------------- ---------------- --------
    /dev/nvme0n1     a34c4f3a0d6f5cec     Linux                                    1         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
    /dev/nvme0n2     a34c4f3a0d6f5cec     Linux                                    2         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
    /dev/nvme1n1     a34c4f3a0d6f5cec     Linux                                    1         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
    /dev/nvme1n2     a34c4f3a0d6f5cec     Linux                                    2         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
    /dev/nvme2n1     a34c4f3a0d6f5cec     Linux                                    1         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
    /dev/nvme2n2     a34c4f3a0d6f5cec     Linux                                    2         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
    /dev/nvme3n1     a34c4f3a0d6f5cec     Linux                                    1         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
    /dev/nvme3n2     a34c4f3a0d6f5cec     Linux                                    2         250.06  GB / 250.06  GB    512   B +  0 B   4.18.0-2
  2. List all connected NVMe subsystems. Check that the command reports them as, for example, nvme0n1 through nvme3n2, and not as, for example, nvme0c0n1 through nvme0c3n1:

    # nvme list-subsys
    
    nvme-subsys0 - NQN=testnqn
    \
     +- nvme0 fc traddr=nn-0x20000090fadd5979:pn-0x10000090fadd5979 host_traddr=nn-0x20000090fac7e1dd:pn-0x10000090fac7e1dd live
     +- nvme1 fc traddr=nn-0x20000090fadd597a:pn-0x10000090fadd597a host_traddr=nn-0x20000090fac7e1dd:pn-0x10000090fac7e1dd live
     +- nvme2 fc traddr=nn-0x20000090fadd5979:pn-0x10000090fadd5979 host_traddr=nn-0x20000090fac7e1de:pn-0x10000090fac7e1de live
     +- nvme3 fc traddr=nn-0x20000090fadd597a:pn-0x10000090fadd597a host_traddr=nn-0x20000090fac7e1de:pn-0x10000090fac7e1de live
    # multipath -ll
    
    mpathae (uuid.8ef20f70-f7d3-4f67-8d84-1bb16b2bfe03) dm-36 NVME,Linux
    size=233G features='1 queue_if_no_path' hwhandler='0' wp=rw
    `-+- policy='service-time 0' prio=50 status=active
      |- 0:1:1:1  nvme0n1 259:0   active ready running
      |- 1:2:1:1  nvme1n1 259:2   active ready running
      |- 2:3:1:1  nvme2n1 259:4   active ready running
      `- 3:4:1:1  nvme3n1 259:6   active ready running
    
    mpathaf (uuid.44c782b4-4e72-4d9e-bc39-c7be0a409f22) dm-39 NVME,Linux
    size=233G features='1 queue_if_no_path' hwhandler='0' wp=rw
    `-+- policy='service-time 0' prio=50 status=active
      |- 0:1:2:2  nvme0n2 259:1   active ready running
      |- 1:2:2:2  nvme1n2 259:3   active ready running
      |- 2:3:2:2  nvme2n2 259:5   active ready running
      `- 3:4:2:2  nvme3n2 259:7   active ready running

Chapter 5. Modifying the DM-Multipath configuration file

By default, DM Multipath provides configuration values for the most common uses of multipathing. In addition, DM Multipath includes support for the most common storage arrays that themselves support DM Multipath. You can override the default configuration values for DM Multipath by editing the /etc/multipath.conf configuration file. If necessary, you can also add an unsupported by default storage array to the configuration file.

For information on the default configuration values, including supported devices, run either of the following commands:

# multipathd show config
# multipath -t
Note

If you run multipath from the initramfs file system and you make any changes to the multipath configuration files, you must rebuild the initramfs file system for the changes to take effect

In the multipath configuration file, you need to specify only the sections that you need for your configuration, or that you wish to change from the default values. If there are sections of the file that are not relevant to your environment or for which you do not need to override the default values, you can leave them commented out, as they are in the initial file.

The configuration file allows regular expression description syntax.

5.1. Configuration file overview

The multipath configuration file is divided into the following sections:

blacklist
Listing of specific devices that will not be considered for multipath.
blacklist_exceptions
Listing of multipath devices that would otherwise be ignored according to the parameters of the blacklist section.
defaults
General default settings for DM Multipath.
multipaths
Settings for the characteristics of individual multipath devices. These values overwrite what is specified in the overrides, devices, and defaults sections of the configuration file.
devices
Settings for the individual storage controllers. These values overwrite what is specified in the defaults section of the configuration file. If you are using a storage array that is not supported by default, you may need to create a devices subsection for your array.
overrides
Settings that are applied to all devices. These values overwrite what is specified in the devices and defaults sections of the configuration file.

When the system determines the attributes of a multipath device, it checks the settings of the separate sections from the multipath.conf file in the following order:

  1. multipaths section
  2. overrides section
  3. devices section
  4. defaults section

5.2. Configuration file defaults

The /etc/multipath.conf configuration file contains a defaults section. This section includes the default configuration of Device Mapper (DM) Multipath. The default values might differ based on your initial device settings.

If you install your machine on a multipath device, the default multipath configuration applies automatically. If you did not set up multipathing during installation, run the following to command to get the default configuration:

# mpathconf --enable

The default configuration includes the following:

# For a complete list of the default configuration values, run either
# multipath -t
# or
# multipathd show config
#
# For a list of configuration options with descriptions, see the multipath.conf
# man page
defaults
{         user_friendly_names yes         find_multipaths yes         enable_foreign \"^$\" }
blacklist_exceptions
{         property \"(SCSI_IDENT_|ID_WWN)\" }

The following table describes the attributes, set in the defaults section of the multipath.conf configuration file. Attributes specified in the multipaths section have higher priority over values in the devices section. Attributes specified in the devices section have higher priority over the default values. Use the overrides section to set attribute values for all device types, even if those device types have a builtin configuration entry in the devices section. The overrides section has no mandatory attributes. However, any attribute set in this section takes precedence over values in the devices or defaults sections.

Table 5.1. Multipath configuration defaults

AttributeDescription

polling_interval

Specifies the interval between two path checks in seconds. For properly functioning paths, the interval between checks gradually increases to max_polling_interval. The default value is 5.

max_polling_interval

Specifies the maximum length of the interval between two path checks in seconds.

The default value is 4 * polling_interval.

find_multipaths

Defines the mode for setting up multipath devices. Available values include:

no: If find_multipaths is set to no, multipath applies rules as with the strict value and the multipathd daemon applies rules as with the greedy value.

yes: If there are at least two devices that are not on the blacklist with the same World Wide Identifier (WWID), or if multipath created a multipath device with a device WWID before (even if that multipath device is no longer present), then the device is treated as a multipath device path.

greedy: Both multipathd and multipath treat every non-blacklisted device as a multipath device path.

smart: Multipath automatically considers that every non-blacklisted device is a multipath device path. If a second path, with the same WWID does not appear within the time set for find_multipaths_timeout, multipath releases the device and enables it for use by the rest of the system. The multipathd daemon applies rules as with the yes value.

strict: This value only treats a device as a multipath path, if you create a multipath device with the device WWID.

The default value is off. The default multipath.conf file sets find_multipaths to yes.

find_multipaths_timeout

This represents the timeout in seconds, to wait for additional paths after detecting the first one, if find_multipaths smart is set. Possible values include:

Positive value: If set with a positive value, the timeout applies for all non-blacklisted devices.

Negative value: If set with a negative value, the timeout applies only to known devices that have an entry in the multipath hardware table, either in the built-in table, or in a device section. Other unknown devices use a timeout of only 1 second to avoid booting delays.

0: The system applies the built-in default for this attribute.

The default value for known hardware is -10. This means that known devices have a 10 second timeout. Unknown devices have a 1 second timeout. If the find_multipaths attribute has a value other than smart, this attribute has no effect.

uxsock_timeout

Set the timeout of multipathd interactive commands in milliseconds.

For systems with a large number of devices, multipathd interactive commands might timeout and fail. If this happens, increase this timeout to resolve the issue.

The default value is 4000.

reassign_maps

Enable reassigning of device-mapper maps. With this option, the multipathd daemon remaps existing device-mapper maps to always point to the multipath device, not the underlying block devices. Possible values are yes and no. The default value is no.

verbosity

The default verbosity value is 2. Higher values increase the verbosity level. Valid levels are between 0 and 4.

path_selector

Specifies the default algorithm to use in determining what path to use for the next I/O operation. Possible values include:

round-robin 0: Loop through every path in the path group, sending the same number of I/O requests, determined by rr_min_io or rr_min_io_rq, to each.

queue-length 0: Send the next group of I/O requests down the path with the least number of outstanding I/O requests.

service-time 0: Send the next group of I/O requests down the path with the shortest estimated service time. This is determined by dividing the total size of the outstanding I/O to each path by the relative throughput.

The default value is service-time 0.

path_grouping_policy

Specifies the default path grouping policy to apply to unspecified multipaths. Possible values include:

failover: 1 path per priority group.

multibus: All valid paths in 1 priority group.

group_by_serial: 1 priority group per detected serial number.

group_by_prio: 1 priority group per path priority value. Priorities are determined by the prio attribute.

group_by_node_name: 1 priority group per target node name. The /sys/class/fc_transport/target*/node_name directory includes target node names.

The default value is failover.

uid_attrs

Set this option to activate merging uevents by WWID. This action might improve uevent processing efficiency. It is also an alternative method to configure the udev properties to use for determining unique path identifiers (WWIDs).

The value of this option is a space separated list of records like type:ATTR, where type is matched against the beginning of the device node name, and ATTR is the name of the udev property to use for matching devices.

If you configure this option and it matches the device node name of a device, it overrides any other configured methods for determining the WWID for this device.

You can enable uevent merging by setting this value to sd:ID_SERIAL dasd:ID_UID nvme:ID_WWN.

The default is unset.

prio

Specifies the default function to call to obtain a path priority value. For example, the ALUA bits in SPC-3 provide an exploitable prio value. Possible values include:

const: Set a priority of 1 to all paths.

emc: Generate the path priority for EMC arrays.

sysfs: Generate the path priority from sysfs. This prioritizer accepts the optional prio_arg value exclusive_pref_bit. The sysfs value uses the sysfs attributes access_state and preferred_path.

alua: Generate the path priority based on the SCSI-3 ALUA settings. If you specify prio alua and prio_args exclusive_pref_bit in your device configuration, multipath creates a path group that contains only the path with the exclusive_pref_bit set, and assigns that path group the highest priority. Refer to the multipath.conf(5) man page for more information about this type of cases.

ontap: Generate the path priority for NetApp arrays.

rdac: Generate the path priority for LSI/Engenio RDAC controller.

hp_sw: Generate the path priority for Compaq/HP controller in active/standby mode.

hds: Generate the path priority for Hitachi HDS Modular storage arrays.

random: Generate a random priority between 1 and 10.

weightedpath: Generate the path priority based on the regular expression and the provided priority as an argument. Requires a prio_args keyword.

path_latency: Generate the path priority based on a latency algorithm. Requires a prio_args keyword.

ana: Generate the path priority based on the NVMe ANA settings. This priority routine is hardware dependent.

datacore: Generate the path priority for some DataCore storage arrays. Requires a prio_args keyword. This priority routine is hardware dependent.

iet: Generate path priority for iSCSI targets based on IP their address. Requires a prio_args keyword. This priority routine is available only with iSCSI.

The default value depends on the detect_prio setting. If detect_prio is set to yes, then the default priority algorithm is sysfs. The only exception is for NetAPP E-Series, where the default is alua. If detect_prio is set to no, the default priority algorithm is const.

prio_args

Arguments to pass to the prio function. This applies only to the following prioritizers:

weighted: Needs a value of the form <hbtl,devname,serial,wwn> <regex1> <prio1> <regex2> <prio2>

hbtl: The Regex value can be of SCSI H:B:T:L format. For example: 1:0:.:. , *:0:0:.

devname: The Regex value can be in device name format. For example: sda, sd.e.

serial: The Regex value can be in serial number format. Look up serial through sysfs, or by running the command multipathd show paths format "%z".

wwn: The Regex value can be in the form host_wwnn:host_wwpn:target_wwnn:target_wwpn. These values can be looked up through sysfs or by running the command multipathd show paths format %N:%R:%n:%r".

path_latency: Requires a value in the form io_num= <integer> base_num=<integer>.

io_num: The number of read IOs, continuously sent to the current path. This value helps calculate the average path latency. Valid values include Integer, [2, 200].

base_num: The base number value of logarithmic scale. This value helps to partition different priority ranks. Valid values include Integer, [2, 10]. The maximum average latency value is 100s and the minimum average latency value is 1us.

alua: If the exclusive_pref_bit value is set, paths with the preferred_path_bit set always create their own path group.

sysfs: If the exclusive_pref_bit value is set, paths with the preferred_path_bit set always create their own path group.

datacore: Requires a value of the form timeout=<milliseconds> preferredsds=<name>.

preferredsds: This value is mandatory and it represents the preferred SDS name.

timeout: This value is optional. Set the timeout for the inquiry in milliseconds.

iet: Requires a value of the form preferredip=<ip_address>.

preferredip: This value is mandatory. This is the preferred IP address, in dotted decimal notation, for iSCSI targets.

The default value is unset.

features

The default extra features of multipath devices, using the format: "number_of_features_plus_arguments feature1 …​".

Possible values for features include:

queue_if_no_path: The same as setting no_path_retry to queue.

pg_init_retries n: Retry path group initialization up to n times before failing. The number must be between 1 and 50.

pg_init_delay_msecs msecs: Number of milliseconds before pg_init retry initiates. The number must be between 0 and 60000.

queue_mode mode: Select the queueing mode per multipath device. The mode value options are bio, rq or mq. These correspond to bio-based, request-based, and block-multiqueue request-based (blk-mq), respectively.

By default, the value is unset. The default can also depend on the kernel parameter dm_mod.use_blk_mq. The two options are mq if it is already set in the parameter, or rq otherwise.

path_checker

Specifies the default method to determine the state of the paths. Possible values include:

readsector0: Read the first sector of the device.

tur: Issue a TEST UNIT READY command to the device.

emc_clariion: Query the EMC Clariion specific EVPD page 0xC0 to determine the path.

hp_sw: Check the path state for HP storage arrays with Active/Standby firmware.

rdac: Check the path state for LSI/Engenio RDAC storage controller.

directio: Read the first sector with direct I/O.

cciss_tur: Check the path state for HP/COMPAQ Smart Array(CCISS) controllers. This is hardware dependent.

none: Does not check the device. Falls back to use values retrieved from sysfs.

The default value is tur.

alias_prefix

This attribute represents the user_friendly_names prefix.

The default value is mpath.

failback

Manages path group failback. Possible values include:

immediate: Specifies immediate failback to the highest priority path group that contains active paths.

manual: Specifies that there is no immediate failback, but that failback can happen only with operator intervention.

followover: Specifies that automatic failback can only be performed when the first path of a path group becomes active. This keeps a node from automatically failing back, when another node requested the failover.

A numeric value greater than zero, specifies deferred failback, and is expressed in seconds.

The default value is manual.

rr_min_io

Specifies the number of I/O requests to route to a path before switching to the next path in the current path group. This setting is only for systems running kernels older than 2.6.31. Newer systems should use rr_min_io_rq. The default value is 1000.

rr_min_io_rq

Specifies the number of I/O requests to route to a path, before switching to the next path in the current path group. Uses a request-based device-mapper-multipath. This setting can be used on systems running current kernels. On systems running kernels older than 2.6.31, use rr_min_io. The default value is 1.

no_path_retry

A numeric value for this attribute specifies the number of times that the path checker must fail for all paths in a multipath device, before disabling queuing.

A value of fail indicates immediate failure, without queuing.

A value of queue indicates that queuing should not stop until the path is fixed.

The default value is fail.

user_friendly_names

Possible values include:

yes: Specifies that the system can use the /etc/multipath/bindings file to assign a persistent and unique alias to the multipath, in the form of mpath<n>.

no: The system uses the WWID as the alias for the multipath. Any device-specific alias you set in the multipaths section of the configuration file, overrides this name.

The default value is no.

queue_without_daemon

If set to no, the multipathd daemon disables queuing for all devices, when it is shut down. The default value is no.

flush_on_last_del

If set to yes, the multipathd daemon disables queuing when the last path to a device is deleted. The default value is no.

max_fds

Sets the maximum number of open file descriptors that can be opened by multipath and the multipathd daemon. This is equivalent to the ulimit -n command. The default value is max, which sets this to the system limit from /proc/sys/fs/nr_open.

checker_timeout

The timeout to use for prioritizers and path checkers that issue SCSI commands with an explicit timeout, in seconds. The sys/block/sd<x>/device/timeout directory contains the default value.

fast_io_fail_tmo

The number of seconds the SCSI layer waits after a problem is detected on an FC remote port, before failing I/O to devices on that remote port. This value must be smaller than the value of dev_loss_tmo. Setting this to off disables the timeout. The default value is 5. The fast_io_fail_tmo option overrides the values of the recovery_tmo and replacement_timeout options of the underlying path devices.

dev_loss_tmo

The number of seconds the SCSI layer waits after a problem is detected on an FC remote port, before removing it from the system. Setting this to infinity will set this to 2147483647 seconds, or 68 years. The OS determines the default value.

eh_deadline

Specifies the maximum number of seconds the SCSI layer spends performing error handling, when SCSI devices fail. After this timeout, the scsi layer performs a full HBA reset. Setting this is necessary in cases where the rport is never lost, so fast_io_fail_tmo and dev_loss_tmo never trigger, but scsi commands still hang. When the SCSI error handler performs the HBA reset, this affects all target paths on that HBA. The eh_deadline value should only be set in cases where all targets on the affected HBAs are multipathed.

The default value is unset.

detect_prio

If this is set to yes, multipath detects if the device is a SCSI device that supports Asymmetric Logical Unit Access (ALUA), or a NVMe device that supports Asymmetric Namespace Access (ANA). If the device supports ALUA, multipath automatically assigns it the alua prioritizer. If the device supports ANA, multipath automatically assigns it the ana prioritizer.

If detect_prio is set to no, or if the device does not support ALUA or ANA, the prio attribute sets the prioritizer.

The default value is yes.

uid_attribute

Specifies the udev attribute to use for the device WWID.

The default value is device dependent: ID_SERIAL for SCSI devices, ID_UID for DASD devices, and ID_WWN for NVMe devices.

force_sync

If set to yes, this parameter prevents path checkers from running in async mode. This means that only one checker runs at a time. This is useful in cases where many multipathd checkers run in parallel, and can cause significant CPU pressure.

The default value is no.

strict_timing

If set to yes, the multipathd daemon starts a new path checker loop after exactly one second, so that each path check occurs at the exactly set seconds for polling_interval. On busy systems, path checks might take longer than one second. The missing ticks are accounted for in the next round. A warning prints if path checks take longer than the set seconds for polling_interval.

The default value is no.

retrigger_tries, retrigger_delay

Use the retrigger_tries and retrigger_delay parameters in conjunction to make multipathd retrigger uevents. If udev fails to completely process the original uevents, this leaves multipath unable to use the device. The retrigger_tries parameter sets the number of times that multipath tries to retrigger a uevent, in case a device is not completely set up. The retrigger_delay parameter sets the number of seconds between retries. Both of these options accept numbers greater than or equal to 0. Setting the retrigger_tries parameter to 0 disables retries. Setting the retrigger_delay parameter to 0 causes the uevent to be reissued on the next loop of the path checker.

The default value of retrigger_tries is 3. The default value of retrigger_delay is 10.

missing_uev_wait_timeout

This attribute controls the number of seconds the multipathd daemon waits to receive a change event from udev for a newly created multipath device. After that it automatically enables device reloads. In most cases, multipathd delays reloads on a device, until it receives a change uevent from the initial table load.

The default value is 30.

deferred_remove

If set to yes, multipathd performs a deferred remove, instead of a regular remove, when the last path device is deleted. This ensures that if a multipathed device is in use when a regular remove is performed and the remove fails, the device is automatically removed, when the last user closes the device. The default value is no.

san_path_err_threshold, san_path_err_forget_rate, san_path_err_recovery_time

If you set all three of these attributes to integers greater than zero, they enable the multipathd daemon to keep shaky paths from reinstating, by monitoring how frequently the path checker fails. If a path checker fails a path more than the value in the san_path_err_threshold attribute, within san_path_err_forget_rate checks, then the multipathd daemon does not reinstate the path until the value of the san_path_err_recovery_time attribute in seconds passes, without any path checker failures.

See the Shaky paths detection section of the multipath.conf(5) for more information.

The default value is no.

marginal_path_double_failed_time, marginal_path_err_sample_time, marginal_path_err_rate_threshold, marginal_path_err_recheck_gap_time

If marginal_path_double_failed_time, marginal_path_err_rate_threshold, and marginal_path_err_recheck_gap_time are set to integers greater than 0 and marginal_path_err_sample_time is set to an integer greater than 120, they enable the multipathd daemon to keep shaky paths from reinstating, by testing the I/O failure rate of paths that repeatedly fail.

If a path fails twice within the value set in the marginal_path_double_failed_time attribute in seconds, the multipathd daemon does not immediately reinstate it, when the path checker determines that it is back up. Instead, multipathd issues a steady stream of read I/Os to the path for the value set in the marginal_path_err_sample_time attribute in seconds. If there are more than the value set in the marginal_path_err_rate_threshold attribute number of errors per thousand I/Os, multipathd waits for marginal_path_err_recheck_gap_time seconds, and then starts another cycle of testing the path with read I/Os. Otherwise, multipathd reinstates the path.

See the Shaky paths detection section of the multipath.conf(5) for more information.

The default value is no.

marginal_pathgroups

Possible values include:

on: When one of the marginal path detecting methods determines that a path is marginal, the system reinstates the path and places it in a separate pathgroup. This group comes into effect only after all the non-marginal path groups are tried first. This prevents the possibility of IO errors occurring while the system can still use some marginal paths. The path returns to a regular path group as soon as it passes monitoring for a configured time.

off: The delay_*_checks, marginal_path_*, and san_path_err_* attributes keep the system from reinstating any marginal, or shaky paths, until they are monitored for a configured time.

fpin: The multipathd daemon receives fpin notifications, sets path states to marginal, and regroups paths, as described for the on value.

The marginal_path_* and san_path_err_* attributes are implicitly set to no.

See the Shaky paths detection section of the multipath.conf(5) for more information.

The default value is no.

log_checker_err

If set to once, multipathd logs the first path checker error at verbosity level 2. The system logs any further errors at verbosity level 3, until the device is restored. If the log_checker_err parameter is set to always, multipathd always logs the path checker error at verbosity level 2. The default value is always.

skip_kpartx

If set to yes, kpartx does not automatically create partitions on the device. This enables you to create a multipath device, without creating partitions, even if the device has a partition table. The default value of this option is no.

max_sectors_kb

Using this option, you can set the max_sectors_kb device queue parameter to the specified value on all underlying paths of a multipath device, before the first activation of a multipath device. Whenever the system creates a new multipath device, the device inherits the max_sectors_kb value from the path devices. Manually raising this value for the multipath device, or lowering this value for the path devices, can cause multipath to create I/O operations larger than the path devices allow. Using the max_sectors_kb parameter is an easy way to set these values, before the creation of a multipath device on top of the path devices, and prevent passing any invalid-sized I/O operations. If you do not set this parameter, the path devices driver sets it automatically, and the multipath device inherits it from the path devices.

ghost_delay

This attribute sets the number of seconds that multipath waits after creating a device with only ghost paths, before marking it ready for use in systemd. This gives the active paths time to appear before the multipath runs the hardware handler to switch the ghost paths to active ones.

Setting this to 0 or no makes multipath immediately mark a device with only ghost paths as ready.

The default value is no.

enable_foreign

This attribute enables or disables foreign libraries.

The value is a regular expression. Foreign libraries are loaded if their name matches the expression.

By default, all libraries are enabled. However, the default configuration file also sets this attribute to "^$", which disables all foreign libraries.

recheck_wwid

If set to yes, when a failed path is restored, the multipathd daemon rechecks the path WWID. If there is a change in the WWID, the path is removed from the current multipath device, and added again as a new path. The multipathd daemon also checks the path WWID again if it is manually re-added.

This option only works for SCSI devices with configuration to use the default uid_attribute, ID_SERIAL, or sysfs, for getting their WWID.

The default value is no.

remove_retries

This option sets the number of times multipath retries removing a device that is in use. Between each attempt, multipath becomes inactive for 1 second. The default value is 0, which means that multipath does not retry the remove.

detect_checker

If set to yes, multipath checks if the device supports ALUA or Redundant Disk Array Controller (RDAC). If the device supports ALUA, multipath assigns it the tur path_checker. If the device supports RDAC, the multipathd daemon assigns it the rdac path_checker. If the device does not support ALUA or RDAC, or the detect_checker is set to no, the path_checker attribute sets the path checker.

The default value is yes.

reservation_key

The mpathpersist parameter uses this service action reservation key. It must be set for all multipath devices using persistent reservations, and it must be the same as the RESERVATION KEY field of the PERSISTENT RESERVE OUT parameter list, which contains an 8-byte value provided by the application client to the device server to identify the I_T nexus. If you use the --param-aptpl option when registering the key with mpathpersist, you must append :aptpl to the end of the reservation key.

This parameter can also be set to file, which causes mpathpersist to automatically store the RESERVATION KEY used to register the multipath device in the prkeys file. The multipathd daemon then uses this key to register additional paths as they appear. When you remove the registration, this automatically removes the RESERVATION KEY from the prkeys file. It is unset by default. If persistent reservations are necessary, it is recommended to set this attribute to file.

all_tg_pt

If this option is set to yes when mpathpersist registers keys, it treats a registered key from one host to one target port, as going from one host to all target ports. This must be set to yes to successfully use mpathpersist on arrays that automatically set and clear registration keys on all target ports from a host, instead of per target port per host. The default value is no.

Additional resources

  • multipath.conf(5) man page

5.3. Configuration file multipaths section

Set attributes of individual multipath devices by using the multipaths section of the multipath.conf configuration file. Device Mapper (DM) Multipath uses these attributes to override all other configuration settings, including those from the overrides section. Refer to Configuration file overrides section for a list of attributes from the overrides section.

The multipaths section recognizes only the multipath subsection as an attribute. The following table shows the attributes that you can set in the multipath subsection, for each specific multipath device. These attributes apply only to one specified multipath. If several multipath subsections match a specific device World Wide Identifier (WWID), the contents of those subsections merge. The settings from latest entries have priority over any previous versions.

Table 5.2. Multipath subsection attributes

AttributeDescription

wwid

Specifies the WWID of the multipath device, to which the multipath attributes apply. This parameter is mandatory for this section of the multipath.conf file.

alias

Specifies the symbolic name for the multipath device, to which the multipath attributes apply. If you are using user_friendly_names, do not set this value to mpath <n>. This might cause conflicts with an automatically assigned user friendly name, and give you incorrect device node names.

The attributes in the following list are optional. If you do not set them, default values from the overrides, devices, or defaults sections apply. Refer to Configuration file defaults for a full description of these attributes.

  • path_grouping_policy
  • path_selector
  • prio
  • prio_args
  • failback
  • no_path_retry
  • rr_min_io
  • rr_min_io_rq
  • flush_on_last_del
  • features
  • reservation_key
  • user_friendly_names
  • deferred_remove
  • san_path_err_threshold
  • san_path_err_forget_rate
  • san_path_err_recovery_time
  • marginal_path_err_sample_time
  • marginal_path_err_rate_threshold
  • marginal_path_err_recheck_gap_time
  • marginal_path_double_failed_time
  • delay_watch_checks
  • delay_wait_checks
  • skip_kpartx
  • max_sectors_kb
  • ghost_delay

The following example shows multipath attributes specified in the configuration file for two specific multipath devices. The first device has a WWID of 3600508b4000156d70001200000b0000 and a symbolic name of yellow.

The second multipath device in the example has a WWID of 1DEC_321816758474 and a symbolic name of red.

Example 5.1. Multipath attributes specification

multipaths {
       multipath {
              wwid                  3600508b4000156d70001200000b0000
              alias                 yellow
              path_grouping_policy  multibus
              path_selector         "round-robin 0"
              failback              manual
              no_path_retry         5
       }
       multipath {
              wwid                  1DEC_321816758474
              alias                 red
        }
}

Additional resources

5.4. Configuration file devices section

Use the devices section of the multipath.conf configuration file to define settings for individual storage controller types. Values set in this section overwrite specified values in the defaults section.

The system identifies the storage controller types by the vendor, product, and revision keywords. These keywords are regular expressions and must match the sysfs information about the specific device.

The devices section recognizes only the device subsection as an attribute. If there are multiple keyword matches for a device, the attributes of all matching entries apply to it. If an attribute is specified in several matching device subsections, later versions of entries have priority over any previous entries.

Important

Configuration attributes in the latest version of the device subsections override attributes in any previous devices subsections and from the defaults section.

The following table shows the attributes that you can set in the device subsection.

Table 5.3. Devices section attributes

AttributeDescription

vendor

Specifies the regular expression to match the device vendor name. This is a mandatory attribute.

product

Specifies the regular expression to match the device product name. This is a mandatory attribute.

revision

Specifies the regular expression to match the device product revision. If the revision attribute is missing, all device revisions match.

product_blacklist

Multipath uses this attribute to create a device blacklist entry that has a vendor attribute that matches the vendor attribute of this device entry, and a product attribute that matches this product_blacklist attribute.

vpd_vendor

Shows the vendor specific Vital Product Data (VPD) page information, using the VPD page abbreviation.

The multipathd daemon uses this information to gather device specific information. Currently only the hp3par VPD page is supported.

hardware_handler

Specifies the hardware handler to use for a particular device type. All possible values are hardware dependent and include:

emc: Hardware handler for DGC class arrays, as CLARiiON CX/AX and EMC VNX and Unity families.

rdac: Hardware handler for LSI/Engenio/NetApp RDAC class, as NetApp SANtricity E/EF Series, and OEM arrays from IBM DELL SGI STK and SUN.

hp_sw: Hardware handler for HP/COMPAQ/DEC HSG80 and MSA/HSV arrays with Active/Standby mode exclusively.

alua: Hardware handler for SCSI-3 ALUA compatible arrays.

ana: Hardware handler for NVMe ANA compatible arrays.

The default value is unset.

Important

​​Linux kernels, versions 4.3 and newer, automatically attach a device handler to known devices. This includes all devices supporting SCSI-3 ALUA). The kernel does not enable changing the handler later on. Setting the hardware_handler attribute for such devices on these kernels takes no effect.

The attributes in the following list are optional. If you do not set them, the default values from the defaults sections apply. Refer to Configuration file defaults for a full description of these attributes.

  • path_grouping_policy
  • uid_attribute
  • getuid_callout
  • path_selector
  • path_checker
  • prio
  • prio_args
  • failback
  • alias_prefix
  • no_path_retry
  • rr_min_io
  • rr_min_io_rq
  • flush_on_last_del
  • features
  • reservation_key
  • user_friendly_names
  • deferred_remove
  • san_path_err_threshold
  • san_path_err_forget_rate
  • san_path_err_recovery_time
  • marginal_path_err_sample_time
  • marginal_path_err_rate_threshold
  • marginal_path_err_recheck_gap_time
  • marginal_path_double_failed_time
  • delay_watch_checks
  • delay_wait_checks
  • skip_kpartx
  • max_sectors_kb
  • ghost_delay
  • all_tg_pt

Additional resources

5.5. Configuration file overrides section

The overrides section recognizes the optional protocol subsection, and can contain multiple protocol subsections. The system matches path devices against the protocol subsection, using the mandatory type attribute. Attributes in a matching protocol subsection have priority over attributes in the rest of the overrides section. If there are multiple matching protocol subsections, later entries have higher priority.

The attributes in the following list are optional. If you do not set them, default values from the devices or defaults sections apply.

  • path_grouping_policy
  • uid_attribute
  • getuid_callout
  • path_selector
  • path_checker
  • alias_prefix
  • features
  • prio
  • prio_args
  • failback
  • no_path_retry
  • rr_min_io
  • rr_min_io_rq
  • flush_on_last_del
  • fast_io_fail_tmo
  • dev_loss_tmo
  • eh_deadline
  • user_friendly_names
  • retain_attached_hw_handler
  • detect_prio
  • detect_checker
  • deferred_remove
  • san_path_err_threshold
  • san_path_err_forget_rate
  • san_path_err_recovery_time
  • marginal_path_err_sample_time
  • marginal_path_err_rate_threshold
  • marginal_path_err_recheck_gap_time
  • marginal_path_double_failed_time
  • delay_watch_checks
  • delay_wait_checks
  • skip_kpartx
  • max_sectors_kb
  • ghost_delay
  • all_tg_pt

The protocol subsection recognizes the following mandatory attribute:

Table 5.4. Protocol subsection attribute

AttributeDescription

type

Specifies the protocol string of the path device. Possible values include:

scsi:fcp, scsi:spi, scsi:ssa, scsi:sbp, scsi:srp, scsi:iscsi, scsi:sas, scsi:adt, scsi:ata, scsi:unspec, ccw, cciss, nvme, undef

This attribute is not a regular expression. The path device protocol string must match exactly.

The attributes in the following list are optional for the protocol subsection. If you do not set them, default values from the overrides, devices or defaults sections apply.

  • fast_io_fail_tmo
  • dev_loss_tmo
  • eh_deadline

Additional resources

5.6. DM Multipath overrides of the device timeout

The recovery_tmo sysfs option controls the timeout for a particular iSCSI device. The following options globally override the recovery_tmo values:

  • The replacement_timeout configuration option globally overrides the recovery_tmo value for all iSCSI devices.
  • For all iSCSI devices that are managed by DM Multipath, the fast_io_fail_tmo option in DM Multipath globally overrides the recovery_tmo value.

    The fast_io_fail_tmo option in DM Multipath also overrides the fast_io_fail_tmo option in Fibre Channel devices.

The DM Multipath fast_io_fail_tmo option takes precedence over replacement_timeout. Red Hat does not recommend using replacement_timeout to override recovery_tmo in devices managed by DM Multipath because DM Multipath always resets recovery_tmo, when the multipathd service reloads.

5.7. Modifying multipath configuration file defaults

The /etc/multipath.conf configuration file includes a defaults section that sets the user_friendly_names parameter to yes, as follows.

defaults {
        user_friendly_names yes
}

This overwrites the default value of the user_friendly_names parameter. The default values that are set in the defaults section on the multipath.conf file, are used by DM Multipath unless they are overwritten by the attributes specified in the devices, multipath, or overrides sections of the multipath.conf file.

Procedure

  1. View the /etc/multipath.conf configuration file, which includes a template of configuration defaults:

    #defaults {
    #       polling_interval        10
    #       path_selector           "round-robin 0"
    #       path_grouping_policy    multibus
    #       uid_attribute           ID_SERIAL
    #       prio                    alua
    #       path_checker            readsector0
    #       rr_min_io               100
    #       max_fds                 8192
    #       rr_weight               priorities
    #       failback                immediate
    #       no_path_retry           fail
    #       user_friendly_names     yes
    #}
  2. Overwrite the default value for any of the configuration parameters. You can copy the relevant line from this template into the defaults section and uncomment it.

    For example, to overwrite the path_grouping_policy parameter to multibus instead of the default value of failover, copy the appropriate line from the template to the initial defaults section of the configuration file, and uncomment it, as follows:

    defaults {
            user_friendly_names     yes
            path_grouping_policy    multibus
    }
  3. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  4. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

Additional resources

  • multipath.conf(5) and multipathd(8) man pages

5.8. Modifying multipath settings for specific devices

In the multipaths section of the multipath.conf configuration file, you can add configurations that are specific to an individual multipath device, referenced by the mandatory WWID parameter.

These defaults are used by DM Multipath and override attributes set in the overrides, defaults, and devices sections of the multipath.conf file. There can be any number of multipath subsections in the multipaths section.

Procedure

  1. Modify the multipaths section for specific multipath device. The following example shows multipath attributes specified in the configuration file for two specific multipath devices:

    • The first device has a WWID of 3600508b4000156d70001200000b0000 and a symbolic name of yellow.
    • The second multipath device in the example has a WWID of 1DEC_321816758474 and a symbolic name of red.

    In this example, the rr_weight attribute is set to priorities.

    multipaths {
           multipath {
                  wwid                  3600508b4000156d70001200000b0000
                  alias                 yellow
                  path_grouping_policy  multibus
                  path_selector         "round-robin 0"
                  failback              manual
                  rr_weight             priorities
                  no_path_retry         5
           }
           multipath {
                  wwid                  1DEC_321816758474
                  alias                 red
                  rr_weight             priorities
            }
    }
  2. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  3. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

Additional resources

  • multipath.conf(5) man page

5.9. Modifying the multipath configuration for specific devices with protocol

You can configure multipath device paths, based on their transport protocol. By using the protocol subsection of the overrides section in the /etc/multipath.conf file, you can override the multipath configuration settings on certain paths. This enables access to multipath devices over multiple transport protocols, like Fiber Channel (FC) or Internet Small Computer Systems Interface (iSCSI).

Options set in the protocol subsection override values in the overrides, devices and defaults sections. These options apply only to devices using a transport protocol which matches the type parameter of the subsection.

Prerequisites

  • You have configured Device Mapper (DM) multipath in your system.
  • You have multipath devices where not all paths use the same transport protocol.

Procedure

  1. View the specific path protocol by running the following:

    # multipathd show paths format "%d %P"
    dev protocol
    sda scsi:ata
    sdb scsi:fcp
    sdc scsi:fcp
  2. Edit the overrides section of the /etc/multipath.conf file, by adding protocol subsections for each multipath type.

    • Settings for path devices, which use the scsi:fcp protocol:

      overrides {
              dev_loss_tmo 60
              fast_io_fail_tmo 8
              protocol {
                      type "scsi:fcp"
                      dev_loss_tmo 70
                      fast_io_fail_tmo 10
                      eh_deadline 360
              }
    • Settings for path devices, which use the scsi:iscsi protocol:

      overrides {
              dev_loss_tmo 60
              fast_io_fail_tmo 8
              protocol {
                      type "scsi:iscsi"
                      dev_loss_tmo 60
                      fast_io_fail_tmo 120
              }
    • Settings for path devices, which use all other protocols:

      overrides {
              dev_loss_tmo 60
              fast_io_fail_tmo 8
              protocol {
                      type "<type of protocol>"
                      dev_loss_tmo 60
                      fast_io_fail_tmo 8
              }

The overrides section can include multiple protocol subsections.

Important

The protocol subsection must include a type parameter. The configuration of all paths with a matching type parameter is then updated with the rest of the parameters listed in the protocol subsection.

Additional resources

  • multipath.conf(5) man page

5.10. Modifying multipath settings for storage controllers

The devices section of the multipath.conf configuration file sets attributes for individual storage devices. These attributes are used by DM Multipath unless they are overwritten by the attributes specified in the multipaths or overrides sections of the multipath.conf file for paths that contain the device. These attributes override the attributes set in the defaults section of the multipath.conf file.

Procedure

  1. View the information on the default configuration value, including supported devices:

    # multipathd show config
    # multipath -t

    Many devices that support multipathing are included by default in a multipath configuration.

  2. Optional: If you need to modify the default configuration values, you can overwrite the default values by including an entry in the configuration file for the device that overwrites those values. You can copy the device configuration defaults for the device that the multipathd show config command displays and override the values that you want to change.
  3. Add a device that is not configured automatically by default to the devices section of the configuration file by setting the vendor and product parameters. Find these values by opening the /sys/block/device_name/device/vendor and /sys/block/device_name/device/model files where device_name is the device to be multipathed, as mentioned in the following example:

    # cat /sys/block/sda/device/vendor
    WINSYS
    # cat /sys/block/sda/device/model
    SF2372
  4. Optional: Specify the additional parameters depending on your specific device:

    active/active device
    Usually there is no need to set additional parameters in this case. If required, you might set path_grouping_policy to multibus. Other parameters you may need to set are no_path_retry and rr_min_io.
    active/passive device
    If it automatically switches paths with I/O to the passive path, you need to change the checker function to one that does not send I/O to the path to test if it is working, otherwise, your device will keep failing over. This means that you have set the path_checker to tur, which works for all SCSI devices that support the Test Unit Ready command, which most do.

    If the device needs a special command to switch paths, then configuring this device for multipath requires a hardware handler kernel module. The current available hardware handler is emc. If this is not sufficient for your device, you might not be able to configure the device for multipath.

    The following example shows a device entry in the multipath configuration file:

    #	}
    #	device {
    #		vendor			"COMPAQ  "
    #		product			"MSA1000         "
    #		path_grouping_policy	multibus
    #		path_checker		tur
    #		rr_weight		priorities
    #	}
    #}
  5. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  6. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

Additional resources

  • multipath.conf(5) and multipathd(8) man pages

5.11. Setting multipath values for all devices

Using the overrides section of the multipath.conf configuration file, you can set a configuration value for all of your devices. This section supports all attributes that are supported by both the devices and defaults section of the multipath.conf configuration file, which is all of the devices section attributes except vendor, product, and revision.

These attributes are used by DM Multipath for all devices unless they are overwritten by the attributes specified in the multipaths section of the multipath.conf file for paths that contain the device. These attributes override the attributes set in the devices and defaults sections of the multipath.conf file.

Procedure

  1. Override device specific settings. For example, you might want all devices to set no_path_retry to fail. Use the following command to turn off queueing, when all paths have failed. This overrides any device specific settings.

    overrides {
            no_path_retry fail
    }
  2. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  3. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

Additional resources

  • multipath.conf(5) man page

Chapter 6. Preventing devices from multipathing

You can configure DM Multipath to ignore selected devices when it configures multipath devices. DM Multipath does not group these ignored devices into a multipath device.

6.1. Conditions when DM Multipath creates a multipath device for a path

DM Multipath has a set of default rules to determine whether to create a multipath device for a path or whether to ignore the path. You can configure the behavior.

If the find_multipaths configuration parameter is set to off, multipath always tries to create a multipath device for every path that is not explicitly disabled. If the find_multipaths configuration parameter is set to on, then multipath creates a device, only if one of following conditions is met:

  • There are at least two paths with the same World-Wide Identification (WWID) that are not disabled.
  • You manually force the creation of the device by specifying a device with the multipath command.
  • A path has the same WWID as a multipath device that was previously created even if that multipath device does not currently exist. Whenever a multipath device is created, multipath remembers the WWID of the device so that it automatically creates the device again as soon as it sees a path with that WWID. This allows you to have multipath automatically choose the correct paths to make into multipath devices, without having to disable multipathing on other devices.

If you have previously created a multipath device without using the find_multipaths parameter and then you later set the parameter to on, you might need to remove the WWIDs of any device you do not want created as a multipath device from the /etc/multipath/wwids file. The following example shows a sample /etc/multipath/wwids file. The WWIDs are enclosed by slashes (/):

# Multipath wwids, Version : 1.0
# NOTE: This file is automatically maintained by multipath and multipathd.
# You should not need to edit this file in normal circumstances.
#
# Valid WWIDs:
/3600d0230000000000e13955cc3757802/
/3600d0230000000000e13955cc3757801/
/3600d0230000000000e13955cc3757800/
/3600d02300069c9ce09d41c31f29d4c00/
/SWINSYS  SF2372         0E13955CC3757802/
/3600d0230000000000e13955cc3757803/

In addition to on and off, you can also set find_multipaths to the following values:

strict
Multipath never accepts paths that have not previously been multipathed and are therefore not in the /etc/multipath/wwids file.
smart
Multipath always accepts non-disabled devices in udev as soon as they appear. If multipathd does not create the device within a timeout set with the find_multipaths_timeout parameter, it will release its claim on the device.

The built-in default value of find_multipaths is off. The default multipath.conf file created by mpathconf, however, will set the value of find_multipaths to on.

When the find_multipaths parameter is set to on, disable multipathing only on the devices with multiple paths that you do not want to be multipathed. Because of this, it will generally not be necessary to disable multipathing on devices.

If you add a previously created multipath device to blacklist, removing the WWID of that device from the /etc/multipath/wwids file by using the -w option can help avoid issues with other programs. For example, to remove the device /dev/sdb with WWID 3600d0230000000000e13954ed5f89300 from the /etc/multipath/wwids file, you can use either of the following methods.

  • Removing a multipath device by using the device name.

    #multipath -w /dev/sdb
    wwid '3600d0230000000000e13954ed5f89300' removed
  • Removing a multipath device by using the WWID of the device.

    #multipath -w 3600d0230000000000e13954ed5f89300
    wwid '3600d0230000000000e13954ed5f89300' removed

You can also use the -W option to update the /etc/multipath/wwids file. This would reset the /etc/multipath/wwids file to only include the WWIDs of the current multipath devices. To reset the file, run the following:

#multipath -W
successfully reset wwids

Additional resources

  • multipath.conf(5) man page

6.2. Criteria for disabling multipathing on certain devices

You can disable multipathing on devices by any of the following criteria:

  • WWID
  • device name
  • device type
  • property
  • protocol

For every device, DM Multipath evaluates these criteria in the following order:

  1. property
  2. devnode
  3. device
  4. protocol
  5. wwid

If a device turns out to be disabled by any of the mentioned criteria, DM Multipath excludes it from handling by multipathd, and does not evaluate the later criteria. For each criteria, the exception list takes precedence over the list of disabled devices, if a device matches both.

Note

By default, a variety of device types are disabled, even after you comment out the initial blacklist section of the configuration file.

6.3. Disabling multipathing by WWID

You can disable multipathing on individual devices by their World-Wide Identification (WWID).

Procedure

  1. Disable devices in the /etc/multipath.conf configuration file using the wwid entry.

    The following example shows the lines in the DM Multipath configuration file that disable a device with a WWID of 26353900f02796769:

    blacklist {
           wwid 26353900f02796769
    }
  2. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  3. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

6.4. Disabling multipathing by device name

You can disable multipathing on device types by device name, so that DM Multipath will not group them into a multipath device.

Procedure

  1. Disable devices in the /etc/multipath.conf configuration file using the devnode entry.

    The following example shows the lines in the DM Multipath configuration file that disable all SCSI devices, because it disables all sd* devices as well:

    blacklist {
           devnode "^sd[a-z]"
    }

    You can use a devnode entry to disable individual devices rather than all devices of a specific type. However, this is not recommended because unless it is statically mapped by udev rules, there is no guarantee that a specific device will have the same name on reboot. For example, a device name could change from /dev/sda to /dev/sdb on reboot.

    By default, DM Multipath disables all devices that are not SCSI, NVMe, or DASD, using the following devnode entry:

    blacklist {
           devnode "!^(sd[a-z]|dasd[a-z]|nvme[0-9])"
    }

    The devices that this entry disables do not generally support DM Multipath.

  2. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  3. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

6.5. Disabling multipathing by device type

You can disable multipathing on devices by using the device section.

Procedure

  1. Disable devices in the /etc/multipath.conf configuration file using the device section.

    The following example disables multipathing on all IBM DS4200 and HP devices:

    blacklist {
           device {
                   vendor  "IBM"
                   product "3S42"       #DS4200 Product 10
           }
           device {
                   vendor  "HP"
                   product ".*"
           }
    }
  2. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  3. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

6.6. Disabling multipathing by udev property

You can disable multipathing on devices by their udev property parameter.

Procedure

  1. Disable devices in the /etc/multipath.conf configuration file using the property parameter. This parameter is a regular expression string that matches against the udev environment variable name for the devices.

    The following example disables multipathing on all devices with the udev property ID_ATA:

    blacklist {
            property "ID_ATA"
    }
  2. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  3. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

6.7. Disabling multipathing by device protocol

You can disable multipathing on devices by using device protocol.

Procedure

  1. Optional: View the protocol that a path is using:

    # multipathd show paths format "%d %P"
  2. Disable devices in the /etc/multipath.conf configuration file using the protocol section.

    The following example disables multipathing on all devices with an undefined protocol or an unknown SCSI transport type:

    blacklist {
            protocol "scsi:unspec"
            protocol "undef"
    }

    DM Multipath recognizes the following protocol strings:

    • scsi:fcp
    • scsi:spi
    • scsi:ssa
    • scsi:sbp
    • scsi:srp
    • scsi:iscsi
    • scsi:sas
    • scsi:adt
    • scsi:ata
    • scsi:unspec
    • ccw
    • cciss
    • nvme
    • undef
  3. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  4. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

6.8. Adding exceptions for devices with disabled multipathing

You can enable multipathing by adding exceptions on devices where multipathing is currently disabled.

Prerequisites

  • Multipathing is disabled on certain devices.

Procedure

  1. Enable multipathing on the devices using the blacklist_exceptions section of the /etc/multipath.conf configuration file.

    When specifying devices in the blacklist_exceptions section of the configuration file, you must specify the exceptions using the same criteria as they were specified in the blacklist section. For example, a WWID exception does not apply to devices disabled by a devnode entry, even if the disabled device is associated with that WWID. Similarly, devnode exceptions apply only to devnode entries, and device exceptions apply only to device entries.

    Example 6.1. An exception by WWID

    If you have a large number of devices and want to multipath only one of them with the WWID of 3600d0230000000000e13955cc3757803, instead of individually disabling each of the devices except the one you want, you could disable all of them, and then enable only the one you want by adding the following lines to the /etc/multipath.conf file:

    blacklist {
            wwid ".*"
    }
    
    blacklist_exceptions {
            wwid "3600d0230000000000e13955cc3757803"
    }

    Alternatively, you could use an exclamation mark (!) to invert the blacklist entry, which disables all devices except the specified WWID:

    blacklist {
            wwid "!3600d0230000000000e13955cc3757803"
    }

    Example 6.2. An exception by udev property

    The property parameter works differently than the other blacklist_exception parameters. The value of the property parameter must match the name of a variable in the udev database. Otherwise, the device is disabled. Using this parameter, you can disable multipathing on certain SCSI devices, such as USB sticks and local hard drives.

    To enable multipathing only on SCSI devices that could reasonably be multipathed, set this parameter to (SCSI_IDENT_|ID_WWN) as in the following example:

    blacklist_exceptions {
            property "(SCSI_IDENT_|ID_WWN)"
    }
  2. Validate the /etc/multipath.conf file after modifying the multipath configuration file by running one of the following commands:

    • To display any configuration errors, run:

      # multipath -t > /dev/null
    • To display the new configuration with the changes added, run:

      # multipath -t
  3. Reload the /etc/multipath.conf file and reconfigure the multipathd daemon for changes to take effect:

    # service multipathd reload

Chapter 7. Managing multipathed volumes

The following are a few commands provided by DM Multipath, which you can use to manage multipath volumes:

  • multipath
  • dmsetup
  • multipathd

7.1. Resizing an online multipath device

If you need to resize an online multipath device, use the following procedure.

Procedure

  1. Resize your physical device.
  2. Execute the following command to find the paths to the logical unit number (LUN):

    # multipath -l
  3. Resize your paths. For SCSI devices, writing a 1 to the rescan file for the device causes the SCSI driver to rescan, as in the following command:

    # echo 1 > /sys/block/path_device/device/rescan

    Ensure that you run this command for each of the path devices. For example, if your path devices are sda, sdb, sde, and sdf, you would run the following commands:

    # echo 1 > /sys/block/sda/device/rescan
    # echo 1 > /sys/block/sdb/device/rescan
    # echo 1 > /sys/block/sde/device/rescan
    # echo 1 > /sys/block/sdf/device/rescan
  4. Resize your multipath device:

    # multipathd resize map multipath_device
  5. Resize the file system (assuming no LVM or DOS partitions are used):

    # resize2fs /dev/mapper/mpatha

7.2. Moving a root file system from a single path device to a multipath device

If you have installed your system on a single-path device and later add another path to the root file system, you will need to move your root file system to a multipathed device. See the following procedure for moving from a single-path to a multipathed device.

Prerequisites

  • You have installed the device-mapper-multipath package.

Procedure

  1. Create the /etc/multipath.conf configuration file, load the multipath module, and enable the multipathd systemd service:

    # yum install device-mapper-multipath
  2. Execute the following command to create the /etc/multipath.conf configuration file, load the multipath module, and set chkconfig for the multipathd to on:

    # mpathconf --enable
  3. If the find_multipaths configuration parameter is not set to yes, edit the blacklist and blacklist_exceptions sections of the /etc/multipath.conf file, as described in Preventing devices from multipathing.
  4. In order for multipath to build a multipath device on top of the root device as soon as it is discovered, enter the following command. This command also ensures that find_multipaths allows the device, even if it only has one path.

    # multipath -a root_devname

    For example, if the root device is /dev/sdb, enter the following command.

    # multipath -a /dev/sdb
    wwid '3600d02300069c9ce09d41c4ac9c53200' added
  5. Confirm that your configuration file is set up correctly by executing the multipath command and search the output for a line of the following format. This indicates that the command failed to create the multipath device.

    date  wwid: ignoring map

    For example, if the WWID of the device is 3600d02300069c9ce09d41c4ac9c53200, you would see a line in the output such as the following:

    # multipath
    Oct 21 09:37:19 | 3600d02300069c9ce09d41c4ac9c53200: ignoring map
  6. Rebuild the initramfs file system with multipath:

    # dracut --force -H --add multipath
  7. Shut the machine down.
  8. Boot the machine.
  9. Make the other paths visible to the machine.

Verification steps

  • Check whether the multipath device is created by running the following command:

    # multipath -l | grep 3600d02300069c9ce09d41c4ac9c53200
    mpatha (3600d02300069c9ce09d41c4ac9c53200) dm-0 3PARdata,VV

7.3. Moving a swap file system from a single path device to a multipath device

By default, swap devices are set up as logical volumes. This does not require any special procedure for configuring them as multipath devices as long as you set up multipathing on the physical volumes that constitute the logical volume group. If your swap device is not an LVM volume, however, and it is mounted by device name, you might need to edit the /etc/fstab file to switch to the appropriate multipath device name.

Procedure

  1. Add the WWID of the device to the /etc/multipath/wwids file:

    # multipath -a swap_devname

    For example, if the root device is /dev/sdb, enter the following command.

    # multipath -a /dev/sdb
    wwid '3600d02300069c9ce09d41c4ac9c53200' added
  2. Confirm that your configuration file is set up correctly by executing the multipath command and search the output for a line of the following format:

    date  wwid: ignoring map

    This indicates that the command failed to create the multipath device.

    For example, if the WWID of the device is 3600d02300069c9ce09d41c4ac9c53200, you would see a line in the output such as the following:

    # multipath
    Oct 21 09:37:19 | 3600d02300069c9ce09d41c4ac9c53200: ignoring map
  3. Set up an alias for the swap device in the /etc/multipath.conf file:

    multipaths {
        multipath {
            wwid WWID_of_swap_device
            alias swapdev
        }
    }
  4. Edit the /etc/fstab file and replace the old device path to the root device with the multipath device.

    For example, if you had the following entry in the /etc/fstab file:

    /dev/sdb2 swap                    swap    defaults        0 0

    Change the entry to the following:

    /dev/mapper/swapdev swap          swap    defaults        0 0
  5. Rebuild the initramfs file system with multipath:

    # dracut --force -H --add multipath
  6. Shut the machine down.
  7. Boot the machine.
  8. Make the other paths visible to the machine.

Verification steps

  • Verify if the swap device is on the multipath device:

    # swapon -s

    For example:

    # swapon -s
    
    Filename                Type          Size Used    Priority
    /dev/dm-3               partition     4169724 0    -2

    The file name should match the multipath swap device.

    # readlink -f /dev/mapper/swapdev
    /dev/dm-3

7.4. Determining device mapper entries with the dmsetup command

You can use the dmsetup command to find out which device mapper entries match the multipathed devices.

Procedure

  • Display all the device mapper devices and their major and minor numbers. The minor numbers determine the name of the dm device. For example, a minor number of 3 corresponds to the multipathed device /dev/dm-3.

    # dmsetup ls
    mpathd  (253:4)
    mpathep1        (253:12)
    mpathfp1        (253:11)
    mpathb  (253:3)
    mpathgp1        (253:14)
    mpathhp1        (253:13)
    mpatha  (253:2)
    mpathh  (253:9)
    mpathg  (253:8)
    VolGroup00-LogVol01     (253:1)
    mpathf  (253:7)
    VolGroup00-LogVol00     (253:0)
    mpathe  (253:6)
    mpathbp1        (253:10)
    mpathd  (253:5)

7.5. Administering the multipathd daemon

The multipathd commands can be used to administer the multipathd daemon.

Procedure

  • View the standard default format for the output of the multipathd show maps command:

    # multipathd show maps
    name sysfs uuid
    mpathc dm-0 360a98000324669436c2b45666c567942
  • Some multipathd commands include a format option followed by a wildcard. Display a list of available wildcards with the following command:

    # multipathd show wildcards
  • Display the multipath devices that multipathd is monitoring, using a format string with multipath wildcards, in regular and raw format:

    list|show maps|multipaths format $format
    list|show maps|multipaths raw format $format

    The multipathd command supports format commands that show the status of multipath devices and paths in "raw" format versions. In raw format, no headers are printed and the fields are not padded to align the columns with the headers. Instead, the fields print exactly as specified in the format string. This output can then be more easily used for scripting. You can display the wildcards used in the format string with the multipathd show wildcards command.

  • Display the paths that multipathd is monitoring, using a format string with multipath wildcards, in regular and raw format:

    list|show paths format $format
    list|show paths raw format $format
  • Display the difference between the non-raw and raw formats for the multipathd show maps. Note that in raw format there are no headers and only a single space between the columns:

    # multipathd show maps format "%n %w %d %s"
    name   uuid                              sysfs vend/prod/rev
    mpathc 360a98000324669436c2b45666c567942 dm-0  NETAPP,LUN
    
    # multipathd show maps raw format "%n %w %d %s"
    mpathc 360a98000324669436c2b45666c567942 dm-0 NETAPP,LUN

Additional resources

  • multipathd(8) man page

Chapter 8. Removing storage devices

You can safely remove a storage device from a running system, which helps prevent system memory overload and data loss.

Prerequisites

  • Before you remove a storage device, you must ensure that you have enough free system memory due to the increased system memory load during an I/O flush. Use the following commands to view the current memory load and free memory of the system:

    # vmstat 1 100
    # free
  • Red Hat does not recommend removing a storage device on a system where:

    • Free memory is less than 5% of the total memory in more than 10 samples per 100.
    • Swapping is active (non-zero si and so columns in the vmstat command output).

8.1. Safe removal of storage devices

Safely removing a storage device from a running system requires a top-to-bottom approach. Start from the top layer, which typically is an application or a file system, and work towards the bottom layer, which is the physical device.

You can use storage devices in multiple ways, and they can have different virtual configurations on top of physical devices. For example, you can group multiple instances of a device into a multipath device, make it part of a RAID, or you can make it part of an LVM group. Additionally, devices can be accessed via a file system, or they can be accessed directly such as a “raw” device.

While using the top-to-bottom approach, you must ensure that:

  • the device that you want to remove is not in use
  • all pending I/O to the device is flushed
  • the operating system is not referencing the storage device

8.2. Removing a block device

You can safely remove a block device from a running system to help prevent system memory overload and data loss.

Warning

Rescanning the SCSI bus or performing any other action that changes the state of the operating system, without following the procedure documented here can cause delays due to I/O timeouts, devices to be removed unexpectedly, or data loss.

Prerequisites

  • If you want to remove a multipath device, and you are unable to access its path devices, disable queueing of the multipath device:

    # multipathd disablequeueing map multipath-device

    This enables the I/O of the device to fail, allowing the applications that are using the device to shut down.

  • Ensure that no other applications or services are using the device that you want to remove.
  • Ensure that you back up the data from the device that you want to remove.

Procedure

  1. Unmount any file systems that are mounted on the device using the umount command.
  2. Remove the device from any MD RAID array or from any LVM volume that it belongs to. Depending on the device type, execute one of the following steps:

    • If the device is a member of an LVM group, and it is a multipath device:

      1. Move the data to another device:

        # pvmove -b /dev/mapper/from-multipath-device /dev/mapper/to-multipath-device
      2. Remove the device from the volume group:

        # vgreduce volume-group /dev/mapper/from-multipath-device
      3. Optional: Remove the LVM metadata from the physical device:

        # pvremove /dev/mapper/from-multipath-device
    • If you are removing a multipath device, execute the following commands:

      1. View all the paths to the device:

        # multipath -l

        The output of this command is required in a later step.

      2. Flush the I/O and remove the multipath device:

        # multipath -f multipath-device
    • If the device is not configured as a multipath device, or if the device is configured as a multipath device and you have previously passed I/O to the individual paths, flush any outstanding I/O to all device paths that are used:

      # blockdev --flushbufs device

      This is important for devices accessed directly where the umount or vgreduce commands do not flush the I/O.

    • If you are removing a SCSI device, execute the following commands:

      1. Remove any reference to the path-based name of the device, such as /dev/sd, /dev/disk/by-path, or the major:minor number, in applications, scripts, or utilities on the system. This ensures that different devices added in the future are not mistaken for the current device.
      2. Remove each path to the device from the SCSI subsystem:

        # echo 1 > /sys/block/device-name/device/delete

        where device-name is retrieved from the output of the multipath -l command, if the device was previously used as a multipath device.

  3. Remove the physical device from a running system. Note that the I/O to other devices does not stop when you remove this device.

Additional resources

  • The multipath(8), pvmove(8), vgreduce(8), blockdev(8) and umount(8) man pages.

Chapter 9. Troubleshooting DM Multipath

If you have trouble implementing a multipath configuration, there are a variety of issues you can check for. The following issues may result in a slow or non-functioning multipath configuration:

The multipath daemon is not running
If you find you have trouble implementing a multipath configuration, ensure that the multipathd daemon is running, as described in Setting up DM Multipath. The multipathd daemon must be running in order to use multipathed devices.
Issues with queue_if_no_path feature
If a multipath device is configured with the features "1 queue_if_no_path" option, then any process that issues I/O hangs until one or more paths are restored.

9.1. Troubleshooting issues with queue_if_no_path feature

If a multipath device is configured with the features "1 queue_if_no_path" option, then any process that issues I/O hangs until one or more paths are restored. To avoid this, set the no_path_retry N parameter in the /etc/multipath.conf file, where N is the number of times the system should retry a path.

If you need to use the features "1 queue_if_no_path" option and you experience the issue noted here, you can disable the queueing policy at runtime for a particular LUN, for which all the paths are unavailable.

Procedure

  • Disable queueing for a specific device:

    # multipathd disablequeueing map device
  • Disable queueing for all devices:

    # multipathd disablequeueing maps

After you have disabled queueing for a device, it will remain disabled until multipathd is restarted or reloaded, or until you execute one of the following commands:

  • Reset queueing to the previous value for a specific device:

    # multipathd restorequeueing map device
  • Reset queueing to the previous value for all devices:

    # multipathd restorequeueing maps

9.2. Troubleshooting with the multipathd interactive console

The multipathd -k command is an interactive interface to the multipathd daemon. Entering this command brings up an interactive multipath console. After executing this command, you can enter help to get a list of available commands and Ctrl+D to quit.

Use the multipathd interactive console to troubleshoot problems you might have with your system.

Procedure

  • Display the multipath configuration, including the defaults, before exiting the console:

    # multipathd -k
    multipathd> show config
    multipathd> Ctrl+D
  • Ensure that multipath has picked up any changes to the multipath.conf file:

    # multipathd -k
    multipathd> reconfigure
    multipathd> Ctrl+D
  • Ensure that the path checker is working properly:

    # multipathd -k
    multipathd> show paths
    multipathd> Ctrl+D
  • You can also run a single multipathd interactive command directly from the command line, without starting the interactive console. For example, to check that multipath has picked up any changes to the multipath.conf file, run:

    # multipathd reconfigure

Chapter 10. Configuring maximum time for storage error recovery with eh_deadline

You can configure the maximum allowed time to recover failed SCSI devices. This configuration guarantees an I/O response time even when storage hardware becomes unresponsive due to a failure.

10.1. The eh_deadline parameter

The SCSI error handling (EH) mechanism attempts to perform error recovery on failed SCSI devices. The SCSI host object eh_deadline parameter enables you to configure the maximum amount of time for the recovery. After the configured time expires, SCSI EH stops and resets the entire host bus adapter (HBA).

Using eh_deadline can reduce the time:

  • to shut off a failed path,
  • to switch a path, or
  • to disable a RAID slice.
Warning

When eh_deadline expires, SCSI EH resets the HBA, which affects all target paths on that HBA, not only the failing one. If some of the redundant paths are not available for other reasons, I/O errors might occur. Enable eh_deadline only if you have a fully redundant multipath configuration on all targets.

The value of the eh_deadline parameter is specified in seconds. The default setting is off, which disables the time limit and allows all of the error recovery to take place.

Scenarios when eh_deadline is useful

In most scenarios, you do not need to enable eh_deadline. Using eh_deadline can be useful in certain specific scenarios. For example if a link loss occurs between a Fibre Channel (FC) switch and a target port, and the HBA does not receive Registered State Change Notifications (RSCNs). In such a case, I/O requests and error recovery commands all time out rather than encounter an error. Setting eh_deadline in this environment puts an upper limit on the recovery time. That enables the failed I/O to be retried on another available path by DM Multipath.

Under the following conditions, the eh_deadline parameter provides no additional benefit, because the I/O and error recovery commands fail immediately, which enables DM Multipath to retry:

  • If RSCNs are enabled
  • If the HBA does not register the link becoming unavailable

10.2. Setting the eh_deadline parameter

This procedure configures the value of the eh_deadline parameter to limit the maximum SCSI recovery time.

Procedure

  • You can configure eh_deadline using either of the following methods:

    • defaults section of the multpath.conf file

      From the defaults section of the multpath.conf file, set the eh_deadline parameter to the required number of seconds:

      # eh_deadline 300
      Note

      From RHEL 8.4, setting the eh_deadline parameter using the defaults section of the multpath.conf file is the preferred method.

      To turn off the eh_deadline parameter with this method, set eh_deadline to off.

    • sysfs

      Write the number of seconds into the /sys/class/scsi_host/host<host-number>/eh_deadline files. For example, to set the eh_deadline parameter through sysfs on SCSI host 6:

      # echo 300 > /sys/class/scsi_host/host6/eh_deadline

      To turn off the eh_deadline parameter with this method, use echo off.

    • Kernel parameter

      Set a default value for all SCSI HBAs using the scsi_mod.eh_deadline kernel parameter.

      # echo 300 > /sys/module/scsi_mod/parameters/eh_deadline

      To turn off the eh_deadline parameter with this method, use echo -1.

Legal Notice

Copyright © 2022 Red Hat, Inc.
The text of and illustrations in this document are licensed by Red Hat under a Creative Commons Attribution–Share Alike 3.0 Unported license ("CC-BY-SA"). An explanation of CC-BY-SA is available at http://creativecommons.org/licenses/by-sa/3.0/. In accordance with CC-BY-SA, if you distribute this document or an adaptation of it, you must provide the URL for the original version.
Red Hat, as the licensor of this document, waives the right to enforce, and agrees not to assert, Section 4d of CC-BY-SA to the fullest extent permitted by applicable law.
Red Hat, Red Hat Enterprise Linux, the Shadowman logo, the Red Hat logo, JBoss, OpenShift, Fedora, the Infinity logo, and RHCE are trademarks of Red Hat, Inc., registered in the United States and other countries.
Linux® is the registered trademark of Linus Torvalds in the United States and other countries.
Java® is a registered trademark of Oracle and/or its affiliates.
XFS® is a trademark of Silicon Graphics International Corp. or its subsidiaries in the United States and/or other countries.
MySQL® is a registered trademark of MySQL AB in the United States, the European Union and other countries.
Node.js® is an official trademark of Joyent. Red Hat is not formally related to or endorsed by the official Joyent Node.js open source or commercial project.
The OpenStack® Word Mark and OpenStack logo are either registered trademarks/service marks or trademarks/service marks of the OpenStack Foundation, in the United States and other countries and are used with the OpenStack Foundation's permission. We are not affiliated with, endorsed or sponsored by the OpenStack Foundation, or the OpenStack community.
All other trademarks are the property of their respective owners.