It looks to be as question 8 describes in the FAQ. "A8. What is close-to-open cache consistency?"
>From what Norm describes, it would seem that NFS would have to have an event driven notify whenever a block is updated on NFS server node. Perhaps OCFS2 or glusterfs might be a better choice? But reading the docs for OCFS2, I am not sure if it operates over the network, but rather two client VMs mount the same partition. I remember when the Oracle guys came to Lugod and said you could do it over ieee 1394. And, iSCSI runs across the network correct? It's not NFS though. On Wed, 2013-10-23 at 06:48 -0400, Ken Bloom wrote: > It should work. See the FAQ at nfs.sourceforge.net, though it sounds > like you're already sounds what it says. > > On Oct 22, 2013 7:15 PM, "Norm Matloff" <matl...@cs.ucdavis.edu> > wrote: > Anyone here know much about mmap()? I have following problem: > > Machines A and B, running Fedora, share files through NFS. I > use mmap() > to create a file-backed array on A and B, shared via the > common file > system. I want changes to the array at A to propagate to B > and vice > versa. > > I can't get this to work consistently, even using msync() > (synchronous > or asynchronous). > > This is for general use software, so using a "better" shared > file system > than NFS is not really an option. > > Any ideas? > > Norm Matloff > > _______________________________________________ > vox-tech mailing list > vox-tech@lists.lugod.org > http://lists.lugod.org/mailman/listinfo/vox-tech > _______________________________________________ > vox-tech mailing list > vox-tech@lists.lugod.org > http://lists.lugod.org/mailman/listinfo/vox-tech _______________________________________________ vox-tech mailing list vox-tech@lists.lugod.org http://lists.lugod.org/mailman/listinfo/vox-tech
