openstack team mailing list archive

[Devin Carlen <devin.carlen@xxxxxxxxx>]

The obvious answer is to have Horizon show whatever nova-client would show here.  However, given that these things now stay around forever, I would argue that the functionality within nova-client itself is suspect and that perhaps there should be a flag to show instances/volumes with an error state. 

Otherwise something that fails just stays around forever, which is certainly not desirable.

Devin 


On Saturday, January 21, 2012 at 10:29 AM, Tres Henry wrote:

> While trying to debug boot-from-volume I've ended up with instances (and volumes for that matter) in various states of failure. Currently Horizon simply displays all instances regardless of state (see attached ss).
> 
> Horizon can filter out instances in an error state very simply, however, that is not necessarily the desired user experience. For example if a user attempts to launch a new instance and the instance goes into an error state before Horizon can render the instances page. In that case if we filtered out instances in an error state the user would never see the instance. It would appear as though nothing happened. 
> 
> This issue exists for any asynchronous process which could result in an object in the system with an unexpected state. This issue also impacts every client (not just Horizon).
> 
> My question to the greater community is this: how should OpenStack present objects in unexpected states consistently so that every client (Horizon, CLI, third-party tool, etc) provides a consistent user experience? 
> _______________________________________________
> Mailing list: https://launchpad.net/~openstack
> Post to : openstack@xxxxxxxxxxxxxxxxxxx (mailto:openstack@xxxxxxxxxxxxxxxxxxx)
> Unsubscribe : https://launchpad.net/~openstack
> More help : https://help.launchpad.net/ListHelp
> 
> 
> 
> 
> Attachments: 
> - error_instances.png
>