On Wed, Apr 20, 2016 at 07:25:00PM -0700, Joe Julian wrote: > > > On 04/20/2016 07:23 PM, Jiffin Tony Thottan wrote: > > > > > >On 21/04/16 04:43, Rick Macklem wrote: > >>Hi, > >> > >>Just to let you know, I did find the email responses to my > >>queries some months ago helpful and I now have a pNFS server > >>for FreeBSD using the GlusterFS port at the alpha test stage. > >>So far I have not made any changes to GlusterFS except the little > >>poll() patch that was already discussed on this list last December. > >> > >>Anyhow, if anyone is interested in taking a look at this, > >>I have a primitive document at: > >> http://people.freebsd.org/~rmacklem/pnfs-setup.txt > >>that will hopefully give you a starting point. > >> > >>Thanks to everyone that helped via email a few months ago, rick > > > >Hi Rick, > > > >Awesome some work man. You have cracked Flexfile layout for gluster > >volume. > > > >I still wondering why you picked knfs instead of nfs-ganesha? There will > >lot of context switches > >between kernel space and user space which may effect the metadata > >performance. > >I still remembering the discussion[1] in which I mentioned to use ganesha > >server as MDS. > >And usually gluster volume won't export using knfs. > > > >-- > >Jiffin > > There's also potential memory deadlocks which is why we stopped doing that > in the first place. Note that Rick is running in FreeBSD and FUSE and their knsfd might not be affected with that. I think it's pretty cool to see FreeBSD with a pNFS server that uses Gluster :) Niels
Attachment:
signature.asc
Description: PGP signature
_______________________________________________ Gluster-devel mailing list Gluster-devel@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-devel
