[Openstack-docs] Instance Resource Quotas : A case study in restructure
Tom Fifield
tom at openstack.org
Sat Jul 13 03:54:57 UTC 2013
Hi all,
I've been having a nice little monologue over at:
https://bugs.launchpad.net/openstack-manuals/+bug/1090654
which is a bug to document the introduction (into grizzly) of a feature
called "Instance Resource Quotas", that allow you to set limits on how
much cpu/disk/network particular instance types consume.
As you can read, I had some initial confusion on where this would go. So
here are my conclusions, provided so we can use it as an excercise in
scoping with the current restructure efforts:
The feature would _not_ be listed in:
* Install guide: the install guide is for a basic description of
installation, and this is not a necessary part
* Config reference: the configuration reference is for parts of the
system that need root access to systems be configured - really, things
that need to be tweaked in {nova,neutron,...}.conf. This feature is
always enabled, and set in user-space, using nova commands.
The feature _may_ be listed in:
* The Operations Guide: it may be a best practice to set instance
resource limits when configuing flavors
* The Secuity Guide: there are security implications of not setting
these limits - where in certain situations guests could eg consume all
the bandwidth, leading to a DoS attack
This feature would predominantely live in:
* The Admin User manual: since configuring flavours lives there, and
this is a part of configuring flavours. A link to hypervisors section in
config ref might be included, with a note on support of the feature
across various hypervisors.
* The `Comprehensive` Admin Manual: if/when this exists, it would
contain an extensive discussion of the "extra_specs" metadata field,
which is used to implement the Instance Resource Quotas
Well that's my take - what's yours? If we can converge on this simple
bug, we have achieved a minor success in scoping :)
Regards,
Tom
More information about the Openstack-docs
mailing list