[dm-devel] IBM (Adaptec) DS400 SAN support - sponsorship

Tue Jul 4 07:12:59 UTC 2006

Kevin Taylor wrote:
> Hello devs,
> 
> I'm hoping that I can find the right person to help me on this list, as
> it seems this list is, by far, the greatest collection of people with
> expertise regarding dm-multipath.
> 
> My desire:
> 
> To locate a person who can write a hardware handler for the DS400 that
> will allow proper active/passive fail-over/fail-back, which we would be
> happy to sponsor financially (for contribution back to the project, of
> course).
> 
> My situation:
> 
> Our company has an existing IBM DS400 SAN that we are trying to deploy
> as part of a Oracle RAC project.  In doing testing and analysis of the
> device and proper configuration of it for this purpose, I have learned
> that in a two controller configuration, this device is active/passive,
> and that in dm-multipath this results in the following configuration
> profile (from "multipath -ll" with all controllers connected):
> 
> ---------------------------------------------
> 
> ora_admin (20000d126733eaf14)
> [size=10 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:3 sde 8:64  [active][faulty]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:1:3 sdk 8:160 [active][ready]
> 
> 20000d126733eaf17
> [size=10 MB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:0 sdb 8:16  [active][faulty]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:1:0 sdh 8:112 [active][ready]
> 
> ora_data1 (20000d126733eaf10)
> [size=200 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:5 sdg 8:96  [active][faulty]
> \_ round-robin 0 [prio=1][enabled]
>  \_ 1:0:1:5 sdm 8:192 [active][ready]
> 
> ora_archive (20000d126733eaf15)
> [size=50 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:4 sdf 8:80  [active][faulty]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:1:4 sdl 8:176 [active][ready]
> 
> ora_redo2 (20000d126733eaf0b)
> [size=68 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:0:2 sdd 8:48  [active][ready]
> \_ round-robin 0 [enabled]
>  \_ 1:0:1:2 sdj 8:144 [active][faulty]
> 
> ora_redo1 (20000d126733eaf0a)
> [size=68 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:0:1 sdc 8:32  [active][ready]
> \_ round-robin 0 [enabled]
>  \_ 1:0:1:1 sdi 8:128 [active][faulty]
> 
> ---------------------------------------------
> 
> And when controller 1 is unplugged, the configuration does the
> following:
> 
> ora_admin (20000d126733eaf14)
> [size=10 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:3 sde 8:64  [active][faulty]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:1:3 sdk 8:160 [active][ready]
> 
> 20000d126733eaf17
> [size=10 MB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:0 sdb 8:16  [active][faulty]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:1:0 sdh 8:112 [active][ready]
> 
> ora_data1 (20000d126733eaf10)
> [size=200 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:5 sdg 8:96  [failed][faulty]
> \_ round-robin 0 [prio=1][enabled]
>  \_ 1:0:1:5 sdm 8:192 [active][ready]
> 
> ora_archive (20000d126733eaf15)
> [size=50 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:4 sdf 8:80  [failed][faulty]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:1:4 sdl 8:176 [active][ready]
> 
> ora_redo2 (20000d126733eaf0b)
> [size=68 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [active]
>  \_ 1:0:0:2 sdd 8:48  [failed][faulty]
> \_ round-robin 0 [enabled]
>  \_ 1:0:1:2 sdj 8:144 [active][faulty]
> 
> ora_redo1 (20000d126733eaf0a)
> [size=68 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:1 sdc 8:32  [failed][faulty]
> \_ round-robin 0 [enabled]
>  \_ 1:0:1:1 sdi 8:128 [failed][faulty]
> 
> --------------------------------------------------
> 
> And when controller 2 is unplugged, the configuration does the
> following:
> 
> ora_admin (20000d126733eaf14)
> [size=10 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:3 sde 8:64  [active][faulty]
> \_ round-robin 0 [active]
>  \_ 1:0:1:3 sdk 8:160 [failed][faulty]
> 
> 20000d126733eaf17
> [size=10 MB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:0 sdb 8:16  [active][faulty]
> \_ round-robin 0 [active]
>  \_ 1:0:1:0 sdh 8:112 [failed][faulty]
> 
> ora_data1 (20000d126733eaf10)
> [size=200 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:5 sdg 8:96  [active][faulty]
> \_ round-robin 0 [enabled]
>  \_ 1:0:1:5 sdm 8:192 [active][faulty]
> 
> ora_archive (20000d126733eaf15)
> [size=50 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [enabled]
>  \_ 1:0:0:4 sdf 8:80  [active][faulty]
> \_ round-robin 0 [active]
>  \_ 1:0:1:4 sdl 8:176 [failed][faulty]
> 
> ora_redo2 (20000d126733eaf0b)
> [size=68 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:0:2 sdd 8:48  [active][ready]
> \_ round-robin 0 [enabled]
>  \_ 1:0:1:2 sdj 8:144 [active][faulty]
> 
> ora_redo1 (20000d126733eaf0a)
> [size=68 GB][features="0"][hwhandler="0"]
> \_ round-robin 0 [prio=1][active]
>  \_ 1:0:0:1 sdc 8:32  [active][ready]
> \_ round-robin 0 [enabled]
>  \_ 1:0:1:1 sdi 8:128 [failed][faulty]
> 
> --------------------------------------------------
> 
> I believe that without a proper hardware handler to instruct the DS400
> to activate the passive controller and shift all LUNS in a timely
> fashion, that this storage unit will not be able to be used in a
> multipath manner for controller failover.  Of course, it's quite
> possible that I am wrong, and there may be another way to make this
> active/passive configuration work without a hardware handler, and I'd be
> very happy to hear about any alternatives anyone may suggest.
> 
It looks as it if this is an RDAC style controller.
If you have the latest sg_utils ready you can try to issue an

sg_inq -e /dev/sgX

If there are any pages listed within the 0xc0 - 0xcf range there is a 
good change this is such a system.
And

sg_inq -e -p=0xc2 /dev/sgX

should lists some meaningful info.

Unfortunately there is currently only limited support for RDAC style 
controllers; you can try using the 'prio_tpc' priority program; this 
will allow you at least to access the storage properly.
To enable proper failover you'll have to ensure to place the storage in 
'AVT' mode (Automatic Volume Transfer); but you'll have to look up in 
the manual on how to do that.
Proper RDAC mode requires an hardware handler; the current 
multipath-tools / device-mapper does not implement this.

> Would anyone here be willing to talk with me about our sponsorship of
> expanding the supported hardware list for dm-multipath to include this
> device?  Does anyone know a person not already on this list who might be
> a good possibility as a developer for such a project?
> 
I have started to implement such an hardware handler, but haven't 
progressed any further due to other tasks.

Drop me a mail if you'd like to support the development.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke			hare at suse.de
SuSE Linux Products GmbH		S390 & zSeries
Maxfeldstraße 5				+49 911 74053 688
90409 Nürnberg				http://www.suse.de