[dm-devel] multipath-tools Makefile

Hannes Reinecke hare at suse.de
Wed Jun 20 06:13:28 UTC 2007


Christophe Varoqui wrote:
> Le mardi 19 juin 2007 à 14:25 -0700, Chandra Seetharaman a écrit :
>> On Tue, 2007-06-19 at 23:15 +0200, Christophe Varoqui wrote:
>>> Le lundi 18 juin 2007 à 17:57 -0700, Chandra Seetharaman a écrit :
>>>> Hi,
>>>>
>>>> Sorry about the late response. I just realized that pp_rdac has been
>>>> added to the multipath tree.
>>>>
>>>> pp_rdac is not needed to support lsi-rdac devices. pp_tpc performs
>>>> equally good. I repeated my tests with pp_tpc and they did work well.
>>>>
>>> rdac has superseded the tpc checker in the upstream tree.
>>>
>> (A):
>>> This checker is currently the only one usable with LSI Engenio hardware
>>> integrated by either IBM or SGI. As such I don't plan on removing it
>>> from the tree, if it's ok with you.
>>>
>> I am little confused here.
>>
>> Background info:
>> ------
>> (1) When I said we do _not_ need pp_rdac, I am referring to path
>> priority callout code(the code I submitted on May 21st). Code under the
>> directory path_priority/pp_rdac.
>>
>> (2) The path checker code I submitted on March 23rd is needed to support
>> the LSI Engenio hardware. This code resides in libcheckers/rdac.c
>> ------
>>
>> (2) is needed and I presume you are referring to that at (A) 
>>
>> (1) is not needed and we can achieve the same with
>> path_priority/pp_tpc/mpath_prio_tpc
>>
> 
> The confusion was on my side, thank you for pointing it.
> I'll get back to my homework then.
> 
Or we finally resolve this issue by renaming pp_tpc to pp_rdac.

pp_tpc is sadly a misnomer as it was originally developed against SGI
TPC arrays, of which I only had insufficient documentation.
Only later I found out that this was actually a RDAC controller.

Christophe, go.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare at suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: Markus Rex, HRB 16746 (AG Nürnberg)




More information about the dm-devel mailing list