Your devices are reported as:
mpath2 (36742b0f0000075500000000000002de4) dm-2 NFINIDAT,InfiniBox
size=954M features='0' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=-1 status=active
`- 2:0:0:3 sdd 8:48 active undef running
Maybe there is a false positive in those fixes.
To continue on this we need very exact description on your device type, storage backend, and anything else that defines your disk setup in HW&SW.
Then depending on availability of such devices or at least a similar setup a reproduction can be tried - if not possible one might suggest different things to be tested on your end, but that is way slower and more error prone.
That change could more be a source of what you describe.
It contains a huge set of backports in regard to ignore "local non-mpath devices"
* Cherry-picked patches to ignore local non-mpath devices add-checker_ timeout- default- config- option. patch params- variable- local.patch th-Fix- possible- string- overflow. patch hwtable- factorization. patch strip-trailing- whitespaces- for-getuid- return- v.patch sysfs_attr- cache.patch setup_thread_ attr-to- uevent. c.patch lists-for- uevent- processing. patch uevent- service- handler- from-main- thread. patch th-rework- sysfs-handling. patch sysfs-device- handling- in-multipathd. patch check-offline- status- for-SCSI- devices. patch for-offline- path-in- get_prio. patch th-Remove- duplicate- calls-to- path_offline. patch dev_loss_ tmo-for- no_path_ retry.patch map-for- device- read-only- setting- changes. patch get-right- sysfs-value- for-checker_ timeout. patch handle- offlined- paths.patch fix-scsi- timeout- code.patch make-tgt_ node_name- work-for- iscsi-devices. patch cleanup- dev_loss_ tmo-issues. patch for-setting- 0-to-fast_ io_fail. patch fast_io_ fail-capping. patch enable- getting- uevents- through- libudev. patch devpath- as-argument- for-sysfs- functions. patch -remove- references- to-sysfs_ device. patch -use-struct- path-as- argument- for-event- pro.patch global- udev-reference- pointer- to-config. patch udev-enumeratio n-during- discovery. patch struct- udev_device- during- discovery. patch debugging- output- when-synchroniz ing-path- states. patch struct- udev_device- instead- of-sysdev. patch Fixup-cciss- discovery. patch udev-devices- during- discovery. patch all-references- to-hand- craftes- sysfs-code. patch libudev- cleanup- and-bugfixes. patch check-if- a-device- belongs- to-multipath. patch and-wwids_ file-multipath. conf-option. patch Check-blacklist s-as-soon- as-possible. patch wwids-file- cleanup- options. patch find_multipaths -option. patch
- 0001-multipath-
- 0002-Make-
- 0003-libmultipa
- 0004-Update-
- 0005-Fixup-
- 0006-Remove-
- 0007-Move-
- 0008-Use-
- 0009-Start-
- 0010-libmultipa
- 0011-Rework-
- 0012-Only-
- 0013-Check-
- 0014-libmultipa
- 0015-Update-
- 0016-Reload-
- 0017-multipath-
- 0018-multipath-
- 0019-multipath-
- 0020-multipath-
- 0021-multipath-
- 0022-Fix-
- 0023-Fix-
- 0024-multipath-
- 0025-Use-
- 0026-multipathd
- 0027-multipathd
- 0028-Add-
- 0029-Use-
- 0030-use-
- 0031-More-
- 0032-Use-
- 0033-discovery-
- 0035-Use-
- 0036-Remove-
- 0037-multipath-
- 0038-multipath-
- 0039-multipath-
- 0040-multipath-
- 0041-add-
- 0042-add-
Your devices are reported as: 550000000000000 2de4) dm-2 NFINIDAT,InfiniBox
mpath2 (36742b0f000007
size=954M features='0' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=-1 status=active
`- 2:0:0:3 sdd 8:48 active undef running
Maybe there is a false positive in those fixes.
To continue on this we need very exact description on your device type, storage backend, and anything else that defines your disk setup in HW&SW.
Then depending on availability of such devices or at least a similar setup a reproduction can be tried - if not possible one might suggest different things to be tested on your end, but that is way slower and more error prone.