Problems with proxmox Multipath with Huawei Oceanstor 5300

[HaunterRed01]

Hello, I currently have this problem in which there is only one fiber link from the storage to a SAN of the DELL blade center, when performing the multipath, it appears that I have two duplicate paths of the same UUID, in the proxmox environment I have only 2 nodes, the biggest problem I am having after configuring my multipath, is that one of the nodes appears in ghost mode.


I shared my .conf:
defaults {
polling_interval 2
path_selector "round-robin 0"
path_grouping_policy multibus
uid_attribute ID_SERIAL
rr_min_io 100
failback inmediate
no_path_retry queue
user_friendly_names yes
}
blacklist {
wwid .*
}

blacklist_exceptions {
wwid "36e472e2100e8d96f3a310ce90000000b"
}

devices {
device {
vendor "HUAWEI"
product "XSG1"
revision "2300"
path_grouping_policy "group_by_prio"
path_checker "tur"
hardware_handler "1 alua"
path_selector "round-robin 0"
prio "alua"
failback "inmediate"
features "0"
rr_weight "uniform"
no_path_retry "72"
rr_min_io_rq 10
rr_min_io 10
user_friendly_names "yes"
fast_io_fail_tmo "5"
}
}
multipaths {
multipath {
wwid "36e472e2100e8d96f3a310ce90000000b"
alias mpath001
}
}

and from the side of huawei storage there is the propertiers of the initiators FC:



Is the same configuration from the other mapping of the blade host

 

[xcientist]

I have the same problem, 4 channels open and 4 duplicate disks

 

[alexskysilk]

the biggest problem I am having after configuring my multipath, is that one of the nodes appears in ghost mode.

Why is this a problem? are you having any issues?

You have a single path, but two controllers. one of them is inactive since there is no path to it you could probably just map the first device directly and ignore the sdc device completely since you would never end up using that path- although this may end up reversed with more luns. there is no harm in operating the way you are.