Updating the devices section of the multipath.conf file

If your host is running RHEL 6.5, SLES 11.3 or an earlier version, you can update the /etc/multipath.conf file. If you are using a later release, simply create an empty /etc/multipath.conf file so the system automatically applies all the default configurations, which includes supported values for NetApp E-Series and EF-Series devices.

The following example shows part of the devices section in the /etc/multipath.conf file. The example shows the vendor ID as NETAPP or LSI and the product ID as INF-01-00. Modify the devices section with product and vendor information to match the configuration of your storage array. If your storage array contains devices from more than one vendor, add additional device blocks with the appropriate attributes and values under the devices section. NetApp has no recommendation on a particular path selector to use. Therefore, the default path selector is selected with the device settings as shown in the example. The command "multipathd show config" shows the path selector in the defaults section.

Note: Only update the devices section of the /etc/multipath.conf file if your host is running RHEL 6.5, SLES 11.3, or an earlier version. For Cluster configurations, set failback to manual as specified in the Interoperability Matrix Tool.
devices { 
  device { 
    vendor                       "(LSI|NETAPP)" 
    product                      "INF-01-00" 
    path_grouping_policy         group_by_prio 
    detect_prio	              yes
    prio                         rdac 
    path_checker                 rdac 
    hardware_handler             "1 rdac" 
    failback                     immediate 
    features                     "2 pg_init_retries 50" 
    no_path_retry                30 
    retain_attached_hw_handler   yes
Note: Other than the parameters listed above, the internal default value for all other parameters is the recommended value and should not be changed. Both the round robin or service time path selectors are fully supported, and NetApp makes no specific recommendation for either one.
Attribute Parameter value Description
path_grouping_policy group_by_prio

The path grouping policy to be applied to this specific vendor and product storage.

detect_prio yes

The system detects the path policy routine.

prio rdac

The program and arguments to determine the path priority routine. The specified routine should return a numeric value specifying the relative priority of this path. Higher numbers have a higher priority.

path_checker rdac

The method used to determine the state of the path.

hardware_handler "1 rdac"

The hardware handler to use for handling device-specific knowledge.

failback immediate

A parameter to tell the daemon how to manage path group failback. In this example, the parameter is set to immediate, so failback occurs in the next check interval in which the path is available. To disable the failback, set this parameter to manual.

When clustering or shared LUN environments are used without the Automatic Load Balancing feature enabled, set this parameter to manual.

features "2 pg_init_retries 50"

Features to be enabled. This parameter sets the kernel parameter pg_init_retries to 50. The pg_init_retries parameter is used to retry the mode select commands.

no_path_retry 30

Specify the number of retries before queuing is disabled. Set this parameter to fail for immediate failure (no queuing). When this parameter is set to queue, queuing continues indefinitely.

The amount of time is equal to the parameter value multiplied by the polling_interval (usually 5), for example, 150 seconds for a no_path_retry value of 30.

retain_attached_hw_handler yes

Specifies that the current hardware handler continues to be used.