[CentOS] IBM Storwize V3700 storage - device names

Fri Nov 8 14:47:39 UTC 2013
Tris Hoar <trishoar at bgfl.org>

On 07/11/2013 21:14, Todor Petkov wrote:
> Hello,
> I have IBM Storwize V3700 storage, connected to 2 IBM x3550 M4 servers
> via fiber channel. The servers are with QLogic ISP2532-based 8Gb Fibre
> Channel to
> PCI Express HBA cards and run Centos 5.10
> When I export a volume to the servers, each of them sees the volume
> twice, i.e  /dev/sdb and /dev/sdc, with the same size.
> Previously I have installed many systems with IBM DS3500 series of
> storage and the servers see one disk per export. I am using the MPP
> drives from this package:
> http://support.netapp.com/NOW/public/apbu/oemcp/apbu_lic.cgi/public/apbu/oemcp/09.03.0C05.0652/rdac-LINUX-09.03.0C05.0652-source.tar.gz
> I came upon the IBM site, saying to configure multipath (I never did it
> for DS3500 series). When I did, a new device came, /dev/dm-7, but my
> goal is to have one /dev/sdX type of device and no device mapper. I read
> that Storwize support DMP RDAC, and DS support MPP RDAC, but does anyone
> else have experience with such setup and can give an advice/hint?
> Thanks in advance.
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> http://lists.centos.org/mailman/listinfo/centos

I's been a while since I set this up, and this on on XIV, not v3700 
(which we also have but it only has VMware connected to it) but this is 
a RHEL 5.10 box, so is reasonably compatible.

This is, IMO, normal behaviour for a multipath device. For example on 
one of our boxes if I run:
[root at server ~]# multipath -ll
mpath0 (20017380011ea0c74) dm-2 IBM,2810XIV
[size=224G][features=1 queue_if_no_path][hwhandler=0][rw]
\_ round-robin 0 [prio=1][active]
  \_ 3:0:0:1 sdb 8:16  [active][ready]
  \_ 3:0:1:1 sdc 8:32  [active][ready]
  \_ 3:0:2:1 sdd 8:48  [active][ready]
  \_ 3:0:3:1 sde 8:64  [active][ready]
  \_ 3:0:4:1 sdf 8:80  [active][ready]
  \_ 3:0:5:1 sdg 8:96  [active][ready]
  \_ 4:0:0:1 sdh 8:112 [active][ready]
  \_ 4:0:1:1 sdi 8:128 [active][ready]
  \_ 4:0:2:1 sdj 8:144 [active][ready]
  \_ 4:0:3:1 sdk 8:160 [active][ready]
  \_ 4:0:4:1 sdl 8:176 [active][ready]
  \_ 4:0:5:1 sdm 8:192 [active][ready]

You can see there are 12 sdx devices, but that all maps to just 1 LUN.
With multipathd installed and running this all maps to a single volume 
under /dev/mpath/mpath0 which I then use LVM to manage:

   --- Physical volume ---
   PV Name               /dev/mpath/mpath0
   VG Name               vg_data
   PV Size               224.00 GB / not usable 4.00 MB
   Allocatable           yes (but full)
   PE Size (KByte)       4096
   Total PE              57343
   Free PE               0
   Allocated PE          57343
   PV UUID               GY4ekC-KuXE-LyW6-kiHB-F9g6-ivB2-BD01Ih

This all works fine and allows us to loose paths to the SAN with out 
disruption to the servers. There is no reason to be using /dev/sdx 
devices to control your underlying hardware, and is in fact coincided 
bad practice as there is no assurance that when the server next boots it 
will detect the hardware in the same order. You should really be using 
UUIDS or device labels to address your storage as that is immutable 
between boots.


This email and any files transmitted with it are confidential
and intended solely for the use of the individual or entity 
to whom they are addressed. If you have received this email 
in error please notify postmaster at bgfl.org

The views expressed within this email are those of the 
individual, and not necessarily those of the organisation