[SOLVED] multipath on iSCSI

[aasami]

Hello all!
I would like to ask for help with configuration of multipath on iSCSI disk in Proxmox 6.2.
I have configured multipath on the server:

[hp12 ~]# iscsiadm -m session
tcp: [1] 10.1.100.112:3260,22 iqn.2000-05.com.3pardata:20220002ac005aab (non-flash)
tcp: [4] 10.1.100.113:3260,121 iqn.2000-05.com.3pardata:21210002ac005aab (non-flash)
tcp: [5] 10.1.100.114:3260,122 iqn.2000-05.com.3pardata:21220002ac005aab (non-flash)
tcp: [6] 10.1.100.111:3260,21 iqn.2000-05.com.3pardata:20210002ac005aab (non-flash)
[hp12 ~]#
[hp12 ~]# multipath -ll
hp3par (360002ac0000000000000001e00005aab) dm-2 3PARdata,VV
size=3.0T features='1 queue_if_no_path' hwhandler='0' wp=rw
`-+- policy='round-robin 0' prio=1 status=active
  |- 2:0:0:0 sdb 8:16 active ready running
  |- 4:0:0:0 sdd 8:48 active ready running
  |- 2:0:1:0 sdc 8:32 active ready running
  `- 4:0:1:0 sde 8:64 active ready running
[hp12 ~]#
[hp12 ~]# dmsetup ls --target multipath
hp3par       (253, 2)
[hp12 ~]#
[hp12 ~]# dmsetup info hp3par
Name:              hp3par
State:             ACTIVE
Read Ahead:        256
Tables present:    LIVE
Open count:        0
Event number:      0
Major, minor:      253, 2
Number of targets: 1
UUID: mpath-360002ac0000000000000001e00005aab

[hp12 ~]#

Now the question is: Do I need to add several iSCSI storages in Proxmox GUI all with different "Portal" IPs (10.1.100.111, 10.1.100.112, 10.1.100.113, 10.1.100.114) to get the desired redundancy?
Or am I completely wrong and need to do something else to configure iSCSI storage with redundant multipath?
There is no coverage on multipath in official docu. I have mostly followed this wiki article.
Thank you for any advice beforehand.

 

[UdoB]

In my understanding you would not tell Proxmox to create (another) iSCSI-Resource. From the top-level Proxmox-GUI that multipath-Device should look and feel like a local disc. The device mapper will supply you by a device /dev/mapper/hp3par which is equivalent to a local disk. The redundancy of multipath is working below that layer. There is no need for PVE to know about that fact.

This is a raw block device. You need to format it. Personally I prefer ZFS, but I am not sure if this is recommended - ZFS has no access to the actual hardware in this constellation...

Just my understanding. I do NOT use a comparable configuration (but would like to...).

 

[aasami]

Thank you @UdoB ! That is exactly what I was hoping to hear. I have removed iSCSI storage in GUI and created new LVM storage on top of existing volume group (VG).

# pvdisplay /dev/mapper/hp3par
  --- Physical volume ---
  PV Name               /dev/mapper/hp3par
  VG Name               vg01
  PV Size               3,00 TiB / not usable 0  
  Allocatable           yes
  PE Size               4,00 MiB
  Total PE              786428
  Free PE               773372
  Allocated PE          13056
  PV UUID               fHkxtx-WXak-txKc-CS0O-uxsk-jpbc-Axg39a

The VG uses multipathed PV /dev/mapper/hp3par exactly what I wanted.
Thank you again my friend.

 

[rcd]

Do you run this from the host or in a vm or container?

 

[aasami]

Do you run this from the host or in a vm or container?

I run it on every physical host in a cluster so I can migrate VMs quickly from node to node.

 

[rcd]

Ok, thanks.