Code review comment for lp:~sandy-walsh/nova/zones3

Revision history for this message
Todd Willey (xtoddx) wrote :

I think this looks pretty good.

Is there any information (RST docs) regarding what meaningful values for capabilities are?

Also, in some flags we are using x=y,a=b syntax for multi-valued flags (defualt_log_levels, etc). Perhaps we can keep the same format? Long term perhaps we could make a new DEFINE_dict or some-such that would parse those into a dict.

It looks like the only information that is being fanned out is the service capabilities, but would it make sense to keep abilities in the database pointing to the services themselves? In other words, why are service capabilities ephemeral with the services starting and stopping? If we keep them in the db we can modify them while running and modify the capabilities with nova-manage. This of course doesn't make sense for things like changing the hypervisor for compute since that would always require a restart, but would it be practical for other capabilities?

review: Needs Information

« Back to merge proposal