Jon K Aimone stated:
> D'oh!
>
> You're correct. The way I phrased that it does. Derived profiles are
> still needed to avoid the proliferation of machine specific profiles
> (potentially thousands in our case). Sorry 'bout that.
>
> Being able to select the target disk based on the pci path could provide
> some determinism regardless of changes in controller enumeration.
We (perfpit) kinda do this. If on a client it fails to find the
target device specified in a client's manifest (since say c6t0d0
changed to c8t0d0) our scripts[1] go back and ask our server to find
the pci path of the disk used in the last install of nevada/osol/s10[2].
It matches up the pci path with the changed disk id (c8t0d0) and then
restarts the install process with the new (c8t0d0) target device in the
client's manifest.
Otherwise the changes in controller enumeration which can and do
change on with just about every new AI image. Turns 'Automated'
into 'All-stop'.
So we'd like to support the support of the pci path :)
Regards,
Sean.
.
[1] we have some scripts to catch this and other potential failures
with AI on a client. Modified AI boot-image.
[2] our server stores an explorer output from every install and a quick
copy of format output.
>
>
> Ethan Quach said the following on 05/20/09 04:32 PM:
>>
>>
>> Jon K Aimone wrote:
>>> Hi,
>>>
>>> In the new target device selection methods it would be convenient to
>>> be able to specify the physical device path. e.g.
>>> /pci at xxx/pci at yyy.... all the way down to the disk slice itself.
>>>
>>> This could be extracted from OBP output of show-disks.
>>>
>>> This would obviate one of the needs for derived profiles.
>>
>> But doesn't it exacerbate another need? -- scalability?
>>
>> Having to store and maintain unique profiles, 1 for each client,
>> I thought was one of the reasons why derived profiles was desired.
>>
>>
>> thanks,
>> -ethan
>>
>>>
>>>
>>> Susan Sohn said the following on 05/20/09 04:11 PM:
>>>> The notes from today's meeting are at:
>>>> http://www.opensolaris.org/os/project/caiman/auto_install/ai_client_redesign_mtg_0520.txt
>>>>
>>>>
>>>>
>>>>
>>>> Proposals for default disk selection and specified disk selection
>>>> are included for review.
>>>>
>>>> In tomorrow's meeting, we will continue to discuss slice/partition
>>>> editing, progress and error reporting,
>>>> as well as any feedback received on today's proposals.
>>>>
>>>> Thanks,
>>>> Sue
>>>>
>>>>
>>>> On 05/18/09 09:02, Susan Sohn wrote:
>>>>> There will be meetings on Tue, Wed, and Thu this week to discuss
>>>>> client redesign.
>>>>> The topic for tomorrow's meeting will be disk selection (default
>>>>> and specified) and
>>>>> slice partitioning. William and Jean's presence is requested,
>>>>> anyone else is welcome.
>>>>>
>>>>> Here are the meeting details:
>>>>>
>>>>> Client redesign meetings:
>>>>> Tue, Wed, Thu 5/19,20,21
>>>>> 8m PT/9am MT/5pm Prague
>>>>>
>>>>> USA Toll-Free: 866-839-8145
>>>>> Caller Paid (Intl): 215-446-3660
>>>>> Participant Code: 9041875
>>>>>
>>>>> Thanks,
>>>>> Sue
>>>>>
>>>>> _______________________________________________
>>>>> caiman-discuss mailing list
>>>>> caiman-discuss at opensolaris.org
>>>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>>>
>>>> _______________________________________________
>>>> caiman-discuss mailing list
>>>> caiman-discuss at opensolaris.org
>>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>>
>>> _______________________________________________
>>> caiman-discuss mailing list
>>> caiman-discuss at opensolaris.org
>>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>>
>
> --
> ~~~~\0/~~~~
> Cheers,
> Jon.
> {-%]
> ========
> If you always do what you've always done, you'll always get what you've
> always gotten.
> - Anon.
> --------
> When someone asks you, "Penny for your thoughts," and you put your two cents
> in, what happens to the other penny?
> - G. Carlin (May 12, 1937 - June 22, 2008)
> ========
>
< sh: line 1: mutt.vcard.filter: not found
< _______________________________________________
< caiman-discuss mailing list
< caiman-discuss at opensolaris.org
< http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
