Shared storage is gone in one host

[Sijjim]

Hello,
I have some problems and need your help, please.
I have 2 exactly same hosts, 2 Dell R620 and Storage MD36xx with extension. Booth hosts configurated identical and use the storage as shared storage in proxmox as an lvm with help of multipath. All works fine a longer time but today local Electricity supplier cutted energy for some hours. They told us before and i shut down all correctly.

After power on again, hardware all come up but one of the 2 machines dont connect to the shared storage. It looks like in config is alle fine but nothing helps. The other machine can see everthing in this but cant boot the vm, previously used by the fault host. In gui -> host1 -> disks i can see all disks, included the ones, dont reachable for booting vm, saved on them but in submenu LVM they not appear. At Shell -> "pvesm status" they appear but marked as "inactive" in fault host, on working they marked as "active".

In attached screenshot pxmx1 is the faulty host, pxmx3 is his working brother. All their vm are von shared storage (mp54 & mp90).


Please, help me the host can use the shared storage again

 

[sb-jw]

It looks like in config is alle fine but nothing helps. The other machine can see everthing in this but cant boot the vm, previously used by the fault host. In gui -> host1 -> disks i can see all disks, included the ones, dont reachable for booting vm, saved on them but in submenu LVM they not appear.

Please do not interpret or relate what you see or think you see. Please tell us everything you checked and what the output was. This is the only way we can help you if we can see for ourselves.

 

[Sijjim]

Hi and thanks for reply. I wrote some returns, please tell me which is really helpful or what you need

pvesm status give at fault host
Name Type Status Total Used Available %
local dir active 277955060 14819016 250968876 5.33%
mp54 lvm inactive 0 0 0 0.00%
mp90 lvm inactive 0 0 0 0.00%
pbs pbs active 1727799936 1639935904 22908 94.91%
pxmx2-main dir disabled 0 0 0 N/A
and on functional host
local dir active 277955060 32913056 232874836 11.84%
mp54 lvm active 5271363584 1254096896 4017266688 23.79%
mp90 lvm active 8789221376 8493465600 295755776 96.64%
pbs pbs active 1727799936 1639935904 22908 94.91%
pxmx2-main dir disabled 0 0 0 N/A

vgscan on fault:
Found volume group "pve" using metadata type lvm2
the other:
Found volume group "mp90" using metadata type lvm2
Found volume group "mp54" using metadata type lvm2
Found volume group "pve" using metadata type lvm2

at pvs, vgs, lvs also on faulty all without the two mp54/90

/etc/multipath.conf is the same in both machines:
GNU nano 7.2 /etc/multipath.conf
blacklist {
wwid .*
}

blacklist_exceptions {
wwid "36d4ae5200098e13300000b8564c1728b"
wwid "36d4ae5200098e13300000b8164c17178"
}

multipaths {
multipath {
wwid "36d4ae5200098e13300000b8564c1728b"
alias mp54
}
multipath {
wwid "36d4ae5200098e13300000b8164c17178"
alias mp90
}
}

defaults {
polling_interval 2
path_selector "round-robin 0"
path_grouping_policy multibus
uid_attribute ID_SERIAL
rr_min_io 100
failback immediate
no_path_retry queue
user_friendly_names yes
}

just as /etc/pve/storage.cfg
dir: local
path /var/lib/vz
content snippets,vztmpl,iso,images,backup,rootdir
shared 0

dir: pxmx2-main
path /mnt/pve/pxmx2-main
content snippets,vztmpl,iso,images,rootdir,backup
is_mountpoint 1
nodes pxmx2

pbs: pbs
datastore extend1
server *removed ip*
content backup
fingerprint xxxxx
prune-backups keep-all=1
username root@pam

lvm: mp54
vgname mp54
content rootdir,images
nodes pxmx1,pxmx3
shared 1

lvm: mp90
vgname mp90
content rootdir,images
nodes pxmx3,pxmx1
shared 1

Can you see an direction, where i find my mistake?

Thanks!