On 7 Feb 2010, at 03:05, Chaskiel Grundman wrote:

> 
> 
> On Sat, 6 Feb 2010, Derrick Brashear wrote:
> 
>> On Sat, Feb 6, 2010 at 4:55 PM, Chaskiel Grundman <[email protected]> 
>> wrote:
>>> 1) Why not RXAFS_WhoAmI? why make the vlserver call the ptserver?
>>> 2) CM's do not interpret ACLs. Clients know their access rights because they
>>> are part of the FetchStatus response.
>> 
>> In disconnected you want to ideally enforce what will be enforced in a
>> replay later to give the best service to the client, but:
> I understand that. Why not cache the CallerAccess or dump the existing 
> axscache (converting pag numbers to something else - hashes of ticket 
> contents perhaps)? Is it because you want to be able to have a user who never 
> stat'd a file before going offline be able to write to it?

Not currently, no. I have no intention of getting the cache manager into the 
business of ACL parsing and enforcement - it just doesn't have enough 
information to be able to make those decisions.

My specific goal here is to handle the case where a machine is powered down 
whilst disconnected. I want to be able to provide the ability for a user to 
access data using the same set of access rights as they had before they 
disconnected. For this, I have the axscache information that's stored in each 
vcache, but that is currently keyed by PAG number - the mappings for which will 
have disappeared when the machine powered down. My intention was to map PAGs to 
viceIds, and to store viceIDs with access information on shutdown. Upon power 
up of a disconnected machine, a super user would then say 'become this viceId', 
and gain that set of access permissions.

Simon.


_______________________________________________
AFS3-standardization mailing list
[email protected]
http://michigan-openafs-lists.central.org/mailman/listinfo/afs3-standardization

Reply via email to