On Wed, Dec 28, 2005 at 07:41:50AM -0600, Jeffry R. Abramson wrote:
> On Sat, 24 Dec 2005 15:42:14 +0100, Axel Thimm <[log in to unmask]> wrote:
> >On Thu, Dec 22, 2005 at 09:22:10PM -0600, Jeffry R. Abramson wrote:
> >> Speaking of SAN's, has anyone gotten the multipath feature in SL42 to work?
> >>  We have an IBM blade equipped with a dual-port Qlogic controller connected
> >> through two independent fabrics to a DS400/Adaptec array with a dual-port
> >> controller.  fdisk -l shows two paths to every logical array, ie. each disk
> >> shows up twice.
> >
> >You should be seeing more than that, you have at least four paths
> >(2xHBA ports x 2xcontroller ports).
> 
> The disk array has a A/B pair of controllers in an active/passive
> configuration.  We have zoned out the passive controller for now so only the
> A controller is used.

Why? If you zone out the passive controller, you'll never have a
failover from one controller to the other.

> sda and sdb are the internal SCSI drives.  sdc through sdf are the two
> volumes on the disk array.  I put a single partition on one of the volumes
> and two on the other so it is easy to see which devices are actually the
> same.  In this case, sdc and sde refer to the same volume as does sdd and
> sdf.  /proc/scsi/scsi is as follows:

> Ran multipath -F followed by multipath -v2 and got the same results. 
> multipath -v3 reveals the following:
> 
> ===== path info sdc (mask 0x1f) =====
> bus = 1
> dev_t = 8:32
> size = 1433354240
> vendor = IBM
> product = DS400    S320
> rev = 7.00
> h:b:t:l = 1:0:0:0
> tgt_node_name = 0x20000000d12671dd
> serial = 00000000002671DD423829CD
> path checker = readsector0 (internal default)

Try

path_grouping_policy    group_by_serial
path_checker            tur

in multipath.conf.


> for x in sda sdb sdc sdd sde sdf ; do /sbin/scsi_id -g -u -s /block/$x; done
> SIBM-ESXSST973401LC____F3LB06LLL00007603NPNN
> SIBM-ESXSST973401LC____F3LB06KLK0000760803PM
> 20000d12671dd29cd
> 20000d12671dd29ce
> 20000d12671dd29cd
> 20000d12671dd29ce

That looks OK.
-- 
Axel.Thimm at ATrpms.net