[OpenAFS-devel] Re: [AFS3-std] Re: Hackathon Summary

[Tom Keiser]

On Fri, Oct 9, 2009 at 3:07 PM, Jeffrey Altman
<jaltman@secure-endpoints.com> wrote:
> Andrew Deason wrote:
>> On Fri, 9 Oct 2009 15:12:14 +0100
>> Simon Wilkinson <sxw@inf.ed.ac.uk> wrote:
>>
>>> Generic Quotas
>>> --------------
>>>
>>> Christof had raised the issue of providing a more generic quota
>>> mechanism, which allows more flexible definitions of what quota might
>>> be (rxosd would like to be able to apply a separate quota to files
>>> under a certain size, for example)
>>>
>>> We discussed implementing this as a set of tag value pairs, with each
>>> pair having a globally defined meaning. Individual tags need not be
>>> implemented on every fileserver - there should be an RPC by which
>>> clients can determine which tags a fileserver supports. We want to
>>> implement this by revising existing RPCs which take quota values, and
>>> use it to replace the quota values that those RPCs already contain.
>>>
>>> Christof will write a document describing this, but we won't block
>>> the RPC refresh on it
>>
>> Upon reading this, I thought, why should we limit this to quotas? Should
>> we have a more generic extensible volume metadata RPC that can be used
>> for any volume-related purpose? I would imagine an xattr-like
>> functionality at the volume level could be potentially useful for
>> numerous things.
>>
> I believe that is what is being described by implementation via a set of
> tag-value pairs where the set of tag names can be queried.
>

As Jeff mentions, the tag-value pairs provide us with considerable
generality (ignoring the TLV/unions debate for now).  One note: we are
making an effort to form a clear distinction between tag-value pairs
that relate to durable metadata, and tag-value pairs that relate to
volatile metadata/server state.  Christof and Hartmut are working on
standardization of the interface that is relevent to on-disk metadata,
and I am responsible for designing the introspection interface for
volatile state.  I believe the rough consensus from Edinburgh is that
tag-value pairs are appropriate for both mechanisms, which is how I'm
planning to proceed for the first draft of the volume state I-D,
barring objections.

-Tom