Derrick Brashear wrote (Mon Jun 08 2009 13:10:26 GMT+0200 (CEST))
On Monday, June 8, 2009, Tom Keiser <[email protected]> wrote:
On Sun, Jun 7, 2009 at 1:51 PM, Matt W. Benjamin<[email protected]> wrote:
Hi,
A priority for openafs is to merge the rxosd changeset. One obstacle to
merging is the use of two 'unused' members in existing structures without prior
coordination. We've been asked to move discussion of the topic into this
forum, which we hereby do.
Hi Matt,
Perhaps I'm missing some context since I couldn't make it to the
workshop this year, but I sense that we're really getting ahead of
ourselves -- rxosd (at least in the 1.4.8 tarball I received at the
Google hackathon) is neither cache coherent, nor consistent. Until
we, as a community, have a chance to re-architect rxosd to follow afs3
semantics, and to properly support transactional semantics with regard
to vnodes and the volume package, I think it's premature to talk about
how afs3 should be modified to deal with these namespace collisions.
Assuming the stated desire to standardize OSD piecemeal, assuming the
assignment of SyncCounter to return information about which protocol
can be used to access a given file affects afs3 consistency in any
manner especially without comment as to what the issue might be
appears specious.
I disbelieve so. Matt's original post did include the complete (I
assume, have not yet verified) list of so far unused/spare fields. A
couple of months ago, Hartmut filed a request for a range of related
RPCs that have already been granted.
Or have I misunderstood the apparent problem?
Cheers
- Felix
_______________________________________________
AFS3-standardization mailing list
[email protected]
http://michigan-openafs-lists.central.org/mailman/listinfo/afs3-standardization
