nova-orchestration team mailing list archive

[Lorin Hochstein <lorin@xxxxxxx>]

On Dec 5, 2011, at 8:30 AM, Sandy Walsh wrote:

> Yeah, tricky. I'm hesitant to do a key/value table like InstanceTypeExtraSpecs since it makes it hard to query. I'd prefer something with a little more structure.
> 

By custom join tables, do you mean adding site-specific tables or fields to the database? For this to work, I think we'd need a simpler mechanism for extending the database than the general sqlalchemy-migrate mechanism, which:

a) is too hard for end-user administrators to customize
b) will not upgrade gracefully because of the sqlalchemy-migrate numbering system



> We may need to simply allow for custom join tables that work with the specific scheduler, but provide some hooks in the ZoneManager to allow the updates to get written when events arrive? Dunno.
> 
> I'm trying to get the base CapacityCache stuff finalized soon, so perhaps that'll be a better discussion point?
> 


I agree, let's hold off on this discussion until there's an implementation of the CapacityCache that we can all poke at with a stick.

Take care,

Lorin
--
Lorin Hochstein, Computer Scientist
USC Information Sciences Institute
703.812.3710
http://www.east.isi.edu/~lorin



> 
> From: Lorin Hochstein [lorin@xxxxxxx]
> Sent: Saturday, December 03, 2011 7:58 PM
> To: Sandy Walsh
> Cc: nova-orchestration@xxxxxxxxxxxxxxxxxxx; nova-scaling@xxxxxxxxxxxxxxxxxxx; dodcs list
> Subject: Re: [Nova-orchestration] Tactical Scheduler/Scalability plan ...
> 
> Hi Sandy:
> 
> Just a note:  we want track usage of additional resources on the compute node(e.g., GPUs, or FPGAs), so we'd like to be able to extend the info tracked in the Capacity Cache to support arbitrary resources.
> 
> What do you think the best way is to support tracking this type of additional data within this scheme?
> 
> Lorin
> --
> Lorin Hochstein, Computer Scientist
> USC Information Sciences Institute
> 703.812.3710
> http://www.east.isi.edu/~lorin
> 
> 
> 
> 
> On Dec 1, 2011, at 6:53 AM, Sandy Walsh wrote:
> 
>> Hi y'all, 
>> 
>> Attached are some notes from the whiteboard noodling myself, comstud and anotherjesse did last week:
>> http://wiki.openstack.org/EssexSchedulerImprovements
>> 
>> It addresses some tactical problems we're facing and, hopefully, sets up a pattern for further improvements and hooks for things like orchestration.
>> 
>> Look forward to your feedback!
>> 
>> Cheers,
>> Sandy
>> 
>> -- 
>> Mailing list: https://launchpad.net/~nova-orchestration
>> Post to     : nova-orchestration@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~nova-orchestration
>> More help   : https://help.launchpad.net/ListHelp
> 
>