On Tue, Jul 14, 2009 at 1:56 AM, Felix Frank<[email protected]> wrote: > Simon Wilkinson wrote (Mon Jul 13 2009 23:39:06 GMT+0200 (CEST)) >> >> On 13 Jul 2009, at 22:26, Alf Wachsmann wrote: >>> >>> I would like to open the discussion about these changes since they >>> will have to be dealt with for the RXOSD changes anyway but provide >>> by themselves already a very useful feature. >> >> Rather than overloading existing fields (especially in the case of >> returning two values bitpacked into MinQuota), I think it's more valuable to >> consider the RPC changes necessary to support this as a first class feature. >> Fortunately, this would appear to be relatively simple. We would need to to >> revise GetVolumeStatus and SetVolumeStatus, and the structures >> AFSFetchVolumeStatus, and AFSStoreVolumeStatus. >> >> I will leave others to speak to the desirability of using the MinQuota >> portion of the volume header - I've got no idea how widespread other uses >> for this field may be. > > Oh it's useful alright - many HEP sites supposedly use AFS as grid and farm > computing storage, and thrashing due to misbehaving parallel jobs (that > create huge numbers of files in single volumes) is a probably common issue. > > Bear in mind that not only are additional RPCs needed, but the volume dump > format must be extended as well. I expect such an extension to be on the > agenda for the upcoming protocol upgrade?
If it's not, well, we should be talking about it. This wouldn't be the only proposal that will require the ability to be backed up! Derrick _______________________________________________ AFS3-standardization mailing list [email protected] http://michigan-openafs-lists.central.org/mailman/listinfo/afs3-standardization
