3

I am setting up multipaths for Veritas backup on SAN storage. I noticed that lsblk shows duplicate disks, which is quite confusing.

For example, both sdc and sdd represent the same disk, and similarly, sde and sdf represent the same device.

sdc                      8:32   0    50G  0 disk
├─sdc1                   8:33   0    50G  0 part
└─san69                253:10   0    50G  0 mpath
  └─san69p1            253:11   0    50G  0 part
sdd                      8:48   0    50G  0 disk
├─sdd1                   8:49   0    50G  0 part
└─san69                253:10   0    50G  0 mpath
  └─san69p1            253:11   0    50G  0 part
sde                      8:64   0    69G  0 disk
├─sde1                   8:65   0    69G  0 part
└─mpathb               253:12   0    69G  0 mpath
  └─mpathb1            253:13   0    69G  0 part  /mnt
sdf                      8:80   0    69G  0 disk
├─sdf1                   8:81   0    69G  0 part
└─mpathb               253:12   0    69G  0 mpath
  └─mpathb1            253:13   0    69G  0 part  /mnt

multipath -ll output is as follow

mpathb (360050763808106804800000000000001) dm-12 IBM,2145
size=69G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
|-+- policy='round-robin 0' prio=50 status=active
| `- 11:0:1:1 sde 8:64 active ready running
`-+- policy='round-robin 0' prio=10 status=enabled
  `- 11:0:3:1 sdf 8:80 active ready running
san69 (360050763808106804800000000000000) dm-10 IBM,2145
size=50G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
|-+- policy='round-robin 0' prio=50 status=enabled
| `- 11:0:3:0 sdc 8:32 active ready running
`-+- policy='round-robin 0' prio=10 status=enabled
  `- 11:0:1:0 sdd 8:48 active ready running
Improve this question
1
  • What is your question actually? If you just want to complain about lsblk showing all block devices, well, that's just the way it is. lsblk just lists everything it sees, so you'll see tons of copies of the same entity, be it dm-multipath or a LVM LV. Yes, it makes lsblk less than useful in any halfway grown-up SAN (one with 100+ LUNs and typically at least 8 paths to each of them, so you end up with sdcx and sdrt and six others all representing the same disk blocks. multipath -ll is your friend, as you already know.
    – TooTea
    May 30 at 15:26

1 Answer 1

Reset to default
3

sdc1, sdd1, sde1, sdf1 are paths to storage device, known as LUN, for a single LUN, you can have many of them (usualy 2 or 4, sometime 8, depending on SAN configuration).(*)

mpathb and san69 are "logical" devices, you build up LVM, or in your case partition on them.(*)

mpathb's data can be access either on sdc1 or sdd1.

multipath driver will take care both of load balancing and failure.

From you data, we can tell you IBM device is connected to SAN with only two ports (see illustration in Switched Fabric on wikipedia)

you sould have /mnt mount over /dev/mapper/mpathb1, that's all you need to worry about.

should a failure occur, you will see

mpathb (360050763808106804800000000000001) dm-12 IBM,2145
size=69G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
|-+- policy='round-robin 0' prio=50 status=active
| `- 11:0:1:1 sde 8:64 active ready running
`-+- policy='round-robin 0' prio=10 status=fail
  `- 11:0:3:1 sdf 8:80 unknown
san69 (360050763808106804800000000000000) dm-10 IBM,2145
size=50G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
|-+- policy='round-robin 0' prio=50 status=fail
| `- 11:0:3:0 sdc 8:32 unknown
`-+- policy='round-robin 0' prio=10 status=enabled
  `- 11:0:1:0 sdd 8:48 active ready running

(I no longer have a multipath device at hand, I make status from memory)

(*) people often use LVM, that is create a physical volume, a volume group above to allow easy increasing of capacity.

Also, if there is only one partition on the disk, often people use the whole disk (same as above, it is easier to increase size).

Improve this answer
0

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .