How do I configure device-mapper-multipath for EMC CLARiion or VNX with ALUA mode in RHEL 6, 7?
Environment
- Red Hat Enterprise Linux 6, 7
device-mapper-multipath(DM-Multipath)- EMC Clariion/VNX/Unity storage array in Active-Active mode 4 (ALUA)
Issue
- How do I configure multipath for EMC CLARiion or VNX for ALUA in RHEL 6?
- I am using an EMC Clariion or VNX storage array and wish to configure multipath in active-active mode 4 (ALUA), what are the recommended configurations for active-active multipath?
- Seeing "emc: ALUA failover mode detected" logged in messages
Resolution
Please see the information within the latest "EMC Host Connectivity Guide for Linux" available from EMC. It is highly recommended to consult EMC for proper configuration based on array class, firmware code, failover mode, etc. Failing to following EMC guidelines can result in the storage configuration being unsupported by EMC and undesired failover behavior during NDU.
-
Per latest Rev A47 P/N 300-003-865 of the
Dell EMC Host Connectivity Guide for Linuxdated December 2018, recommends the following clauses all Clariion/VNX/Unity class arrays that support ALUA:-
RHEL 6.0 - 6.2 (see pg 197-198 within above vendor guide)
devices { device { vendor "DGC" product ".*" product_blacklist "LUNZ" : path_checker emc_clariion ### Rev 47 alua hardware_handler "1 alua" ### modified for alua prio alua ### modified for alua : } } -
RHEL 6.3 and later (see pg 198 within above vendor guide)
devices { device { vendor "DGC" product ".*" product_blacklist "LUNZ" : path_checker emc_clariion ### Rev 47 alua hardware_handler "1 alua" ### modified for alua prio tpg_pref ### modified for alua : } } -
RHEL7.0 and later (see pg 199-200 within above vendor guide)
devices { device { vendor "DGC" product ".*" product_blacklist "LUNZ" : path_checker emc_clariion ### Rev 47 alua hardware_handler "1 alua" ### modified for alua prio alua ### modified for alua : } }
NOTE! The above only serves to highlight which fields within the device clause are to be modified for ALUA support. Do not simply cut & paste the above into multipath without the rest of the required fields.
-
-
Add and/or update the devices clause for DGC (CLARiiON) or VNX storage arrays within multipath.conf per above EMC document.
- if a custom DGC clause already exists within
multipath.conffile:- modify the three fields per above within the existing DGC device clause.
- if no DGC clause is currently present within
multipath.conffile:- place a copy of the default device clause for DGC (CLARiiON) or VNX storage arrays within multipath.conf
- a copy of the default device clause is available from
/usr/share/doc/device-mapper-multipath-0.4.9/multipath.conf.defaults - place the device { ... } clause between devices { ... } as show above to maintain proper multipath.conf syntax
- a copy of the default device clause is available from
- remove the leading '#' comment character from the copied device clause
- modify the three fields per above within the newly copied DGC device clause within
multipath.conf
- place a copy of the default device clause for DGC (CLARiiON) or VNX storage arrays within multipath.conf
- if a custom DGC clause already exists within
-
Reload the
multipathdservice for changes to take effect:
For RHEL 6:
# service multipathd reload
For RHEL 7:
# systemctl reload multipathd
- For more details on verifying the configuration, refer to Diagnostic Steps from Configuring device-mapper-multipath for an EMC CLARiiON configured in ALUA mode with RHEL 5
- You should update your initramfs boot image file in order to ensure multipath.conf changes are picked up at boot time. See Why are changes I made to /etc/multipath.conf not taking effect on boot on my boot-from-multipath RHEL system? for specific steps.
- For additional information, see How to rebuild the initial ramdisk image in Red Hat Enterprise Linux
- You should update your initramfs boot image file in order to ensure multipath.conf changes are picked up at boot time. See Why are changes I made to /etc/multipath.conf not taking effect on boot on my boot-from-multipath RHEL system? for specific steps.
Root Cause
- The default
path_checkerisemc_clariion. As per EMC,emc_clariionpath checker is optimized for Passive Not Ready (PNR) failover mode 1 and not for ALUA failover mode 4. - The hardware handler for PNR configurations is "emc", the hardware handler will output "emc: ALUA failover mode detected" when it detects ALUA mode capability within storage. This is a good indicator that the multipath.conf configuration is not correct for the storage configuration.
- See How do I tell if ALUA is enabled on a LUN? for steps on how to confirm storage is configured in ALUA mode.
This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.