One other small change we may consider is being able to set mode flags on a
compound, to tell the server what parts of A-C-I-D we want, if any.
Cheers,
Erez.
--
Find and fix application performance issues faster with
...@cs.stonybrook.edu, "Garima Gehlot"
> <ggeh...@cs.stonybrook.edu>, "Jasmit Saluja" <jsal...@cs.stonybrook.edu>
> Sent: Friday, April 8, 2016 6:01:27 PM
> Subject: Re: [Nfs-ganesha-devel] NFS Transactional Compound Project
>
> On Fri, Apr 8, 20
On Fri, Apr 8, 2016 at 5:21 PM, J. Bruce Fields wrote:
> On Wed, Apr 06, 2016 at 10:37:51AM -0400, Ming Chen wrote:
>> We have just created a github repo about our transactional NFS
>> compounds project: https://github.com/sbu-fsl/txn-compound
>
> To me a "transaction" is an
On Wed, Apr 06, 2016 at 10:37:51AM -0400, Ming Chen wrote:
> We have just created a github repo about our transactional NFS
> compounds project: https://github.com/sbu-fsl/txn-compound
To me a "transaction" is an atomic operation that succeeds or fails as a
unit, which an NFSv4 compound isn't.
I
Thanks. Ming, can we look at sqlite too?
Cheers,
Erez.
> On Apr 6, 2016, at 4:01 PM, Geoff Kuenning wrote:
>
> One thing that leaps to mind is sqlite. More generally, anything that makes
> heavy use of fsync probably would benefit from transactional compounds. I'm
>
One thing that leaps to mind is sqlite. More generally, anything
that makes heavy use of fsync probably would benefit from
transactional compounds. I'm particularly thinking of the results
in the iBench paper.
> Hi there,
>
> We have just created a github repo about our transactional NFS
>
Hi,
As we've discussed previously off-list, the general capability for
library-based nfsv4(1) client access has a large number of potential
applications all by itself. Thanks very much for making this available.
As the foregoing implies, I'm personally very interested in the potential for