On Fri, Jul 22, 2016 at 08:29:39AM +1100, Blair Bethwaite wrote: > Ken, Ira, John - > > Thanks a lot for the replies. Our initial setup is simply running > samba atop a cephfs kernel mount, and initial cursory checks seem to > show the basics are working as expected (even clustered with ctdb - > what are your concerns here Ira?). Though we've yet to try any of our > planned test scenarios/datasets. > > From the conversation here I'm thinking we'd be better off using > Xenial for our CephFS+Samba test nodes at the moment... though I only > see packages for Precise and Trusty on gitbuilder.ceph.com at the > moment. We were planning to compare samba on kernel mount versus ceph > vfs. > > Can someone clarify what state the cephfs kernel client in RHEL 7.x > will be in when RHCS 2.0 is released, i.e., will that be back-ported > or are RHEL users expected to use FUSE? (I'm happy to go ask support > directly but I suspect this is useful information for others too). Please note: icooper@xxxxxxxxxx is not on this list, so I can't reply from there ;). I recommend you test your setup THROUGHLY. CTDB is really in place to handle failures, so test node failures. I recommend hard power-offs for this. The real world is rarely as kind as a "nice" poweroff, and it might trick you into thinking more is working than is. If you get it right, all the nodes should end up banned if I remember right. It's been a bit since I tested with stock settings. For my setup I use: FUSE for the mount so CTDB works. vfs_ceph for the Samba datapath. That's what I'd recommend based on talking to people. No idea on RHCS/Kernel. Cheers, -Ira -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
