Re: [openstack-dev] [neutron][trunk] objects within transactions

[Ihar Hrachyshka]

Gary Kotton  wrote:


Hi,
It is unclear what our policy is with objects. My understanding is that  
an object should handle all of the database interaction. Is that not one  
of the benefits of the objects? I  
postedhttps://review.openstack.org/#/c/365459 and it has a number of  
interesting comments – the main is that drivers are using the PRECOMIT  
callbacks. Can someone please point out to where these are use. That  
would at least provide some justification for this being blocked.

Thanks


Objects indeed provide some transactional guarantees, but it does not mean  
that consuming code may not have any additional transactional requirements  
that will need to be fulfilled by some other code. In the case you  
mentioned, transaction is needed to guarantee that all event consumers have  
a chance to handle update payload before we actually hit database with our  
update).


Ihar

__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Re: [openstack-dev] [neutron][trunk] objects within transactions

[Gary Kotton]

Thanks!


On 9/5/16, 6:20 PM, "Jakub Libosvar"  wrote:

On 05/09/16 16:45, Gary Kotton wrote:
> Hi,
>
> It is unclear what our policy is with objects. My understanding is that
> an object should handle all of the database interaction. Is that not one
> of the benefits of the objects? I posted
> https://review.openstack.org/#/c/365459 and it has a number of
> interesting comments – the main is that drivers are using the PRECOMIT
> callbacks. Can someone please point out to where these are use. That
> would at least provide some justification for this being blocked.
>
> Thanks
>
> Gary
As per comments on the change, they are used in ODL [1].

Kuba

[1] 

https://review.openstack.org/#/c/342459/3/networking_odl/trunk/driver_v2.py@53
>
>
>
> __
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>


__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev



__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Re: [openstack-dev] [neutron][trunk] objects within transactions

[Jakub Libosvar]

On 05/09/16 16:45, Gary Kotton wrote:

Hi,

It is unclear what our policy is with objects. My understanding is that
an object should handle all of the database interaction. Is that not one
of the benefits of the objects? I posted
https://review.openstack.org/#/c/365459 and it has a number of
interesting comments – the main is that drivers are using the PRECOMIT
callbacks. Can someone please point out to where these are use. That
would at least provide some justification for this being blocked.

Thanks

Gary

As per comments on the change, they are used in ODL [1].

Kuba

[1] 
https://review.openstack.org/#/c/342459/3/networking_odl/trunk/driver_v2.py@53




__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev




__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev