Re: [openstack-dev] [Heat] Question regarding Stack updates and templates

2014-09-25 Thread Zane Bitter

On 22/09/14 11:04, Anant Patil wrote:

Hi,

In convergence, we discuss about having concurrent updates to a stack. I
wanted to know if it is safe to assume that the an update will be a
super set of it's previous updates. Understanding this is critical to
arrive at implementation of concurrent stack operations.

Assuming that an admin will have VCS setup and will issue requests by
checking-out the template and modifying it, I could see that the updates
will be incremental and not discreet. Is this assumption correct? When
an update is issued before a previous update is complete, would the
template for that be based on the template of previously issued
incomplete update or the last completed one?


Neither.

The only thing that matters is that we get to the state described in the 
latest template. With multiple updates rippling through at a time, it's 
likely that neither the last completed update nor the last issued update 
is representative of the current state that we need to modify.


Actually, given that updates can fail part-way through, that's already 
the case. That's why we now update the template incrementally in the 
database as we make individual changes.


cheers,
Zane.

___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


[openstack-dev] [Heat] Question regarding Stack updates and templates

2014-09-22 Thread Anant Patil
Hi,

In convergence, we discuss about having concurrent updates to a stack. I
wanted to know if it is safe to assume that the an update will be a
super set of it's previous updates. Understanding this is critical to
arrive at implementation of concurrent stack operations.

Assuming that an admin will have VCS setup and will issue requests by
checking-out the template and modifying it, I could see that the updates
will be incremental and not discreet. Is this assumption correct? When
an update is issued before a previous update is complete, would the
template for that be based on the template of previously issued
incomplete update or the last completed one?

- Anant

___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


Re: [openstack-dev] [Heat] Question regarding Stack updates and templates

2014-09-22 Thread Angus Salkeld
On Tue, Sep 23, 2014 at 1:04 AM, Anant Patil anant.pa...@hp.com wrote:

 Hi,

 In convergence, we discuss about having concurrent updates to a stack. I
 wanted to know if it is safe to assume that the an update will be a
 super set of it's previous updates. Understanding this is critical to
 arrive at implementation of concurrent stack operations.

 Assuming that an admin will have VCS setup and will issue requests by
 checking-out the template and modifying it, I could see that the updates
 will be incremental and not discreet. Is this assumption correct? When
 an update is issued before a previous update is complete, would the
 template for that be based on the template of previously issued
 incomplete update or the last completed one?


I don't think you can assume anything about the update. What if the user
just posts a
totally different template? That is still a valid update. Or they post an
empty template
to delete the resources.

-Angus



 - Anant

 ___
 OpenStack-dev mailing list
 OpenStack-dev@lists.openstack.org
 http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


Re: [openstack-dev] [Heat] Question regarding Stack updates and templates

2014-09-22 Thread Clint Byrum
Excerpts from Angus Salkeld's message of 2014-09-22 20:31:46 -0700:
 On Tue, Sep 23, 2014 at 1:04 AM, Anant Patil anant.pa...@hp.com wrote:
 
  Hi,
 
  In convergence, we discuss about having concurrent updates to a stack. I
  wanted to know if it is safe to assume that the an update will be a
  super set of it's previous updates. Understanding this is critical to
  arrive at implementation of concurrent stack operations.
 
  Assuming that an admin will have VCS setup and will issue requests by
  checking-out the template and modifying it, I could see that the updates
  will be incremental and not discreet. Is this assumption correct? When
  an update is issued before a previous update is complete, would the
  template for that be based on the template of previously issued
  incomplete update or the last completed one?
 
 
 I don't think you can assume anything about the update. What if the user
 just posts a
 totally different template? That is still a valid update. Or they post an
 empty template
 to delete the resources.

Agreed. The new template simply replaces the graph with a new version.
If that graph happens to change everything, then the old things will now
be in a should not exist desired state and the new template convergence
should remove them when it gets to the garbage collection at the end.

___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev