Unity: BFS RHEL8.3 server with DM-MPIO does not boot when Unity LUN is inactive

Red Hat server boot from SAN (BFS) does not boot when Unity LUN is in an Inactive state.

On Red Hat servers (both BFS and Local Boot), the fdisk -l command does not complete when the LUN is set to Inactive.

On Red Hat servers (both BFS and Local Boot), the emcgrab, and sosreport commands do not complete when LUNs is set to Inactive.

Red Hat servers report I/O errors for Inactive LUNs on the console screen and in the messages as shown below:

blk_update_request: I/O error, dev sdh, sector 20971392 op 0x0:(READ) flags 0x4200 phys_seg 1 prio class 0
Buffer I/O error on sdh, logical block 2621424, async page read

The multipath -ll command shows that the path is not configured correctly as below:

mpathb (360xxxxxxxxxxxxxxxxxxxxxxxxx) dm-1 DGC,VRAID
size=10G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=23 status=enabled
| |- 6:0:0:1 sdb 8:16  failed ready running
| |- 7:0:0:1 sdf 8:80  failed ready running
| `- 7:0:1:1 sdh 8:112 failed ready running
`-+- policy='service-time 0' prio=50 status=enabled
  `- 6:0:1:1 sdd 8:48  failed ready running

Red Hat Enterprise Linux 7.4 introduces detect_checker as a stanza parameter whose default value is yes. The Multipath I/O (MPIO) detects if a device supports Asymmetric Logical Unit Access (ALUA). If it supports ALUA and the detect_checker value is yes, the MPIO overrides the configured path_checker and uses the TUR checker instead and the user-defined path_checker emc_clariion is ignored. The TUR checker does not properly handle SCSI sense from Unity Inactive LUNs. DM-MPIO repeatedly restores the paths.

If a Red Hat server is booted from SAN, multipathd startup in the initramfs phase does not complete so the operating system cannot go to the mount real rootFS phase.

If you Local Boot a Red Hat server, the fdisk -l command stops responding because I/O to the inactive LUNs are not completed.

Make DM-MPIO use emc_clariion_checker as it can handle the inactive LUN state properly.

Back up the original /etc/multpath.conf file and modify it as shown below.

See the latest Host Connectivity Guide for Linux and the section for Red Hat Enterprise Linux 7.4 and later for /etc/multpath.conf.

defaults {
        user_friendly_names yes
}
devices {
        device {
                vendor "DGC"
                product ".*"
                product_blacklist "LUNZ"
                path_grouping_policy "group_by_prio"
                path_checker "emc_clariion"
                features "1 queue_if_no_path"
                hardware_handler "1 emc"
                prio "emc"
                failback immediate
                rr_weight "uniform"
                no_path_retry 60
                retain_attached_hw_handler yes
                detect_prio yes
                detect_checker no
        }
}

Below is Boot from SAN (BFS) Red Hat server only.
If your BFS Red Hat server is experiencing boot issues, change the LUN states for the server from inactive to active. Then the operating system completes the boot process. After the operating system boots, modify the /etc/multpath.conf file described as above. After editing the /etc/multpath.conf file, rebuild initramfs multipath with the dracut command. It is recommended to create a backup of an original initramfs file before rebuilding initramfs.

dracut --force --add multipath

For more details for rebuilding initramfs, consult Red Hat support.