On Tue, 29 Mar 2016 20:45:38 -0400 (EDT) Jason Dillaman wrote: > Very good points w.r.t. easing migration. There is no plan to remove > support for image format 1 -- it's more about being able to concentrate > our attention to format 2 and to nudge users towards creating new images > in format 2. > The release notes you cited earlier https://github.com/ceph/ceph/blob/master/doc/release-notes.rst state TWICE (mistake obviously) that format 1 is is going bye-bye: --- * The rbd legacy image format (version 1) is deprecated with the Jewel release. Attempting to create a new version 1 RBD image will result in a warning. Future releases of Ceph will remove support for version 1 RBD images. --- Also while I tend to read everything and in sequence, the format of the release notes can be very misleading. As in people who are looking just at the "Upgrading from Hammer" bits will miss all the very vital information in the Infernalis upgrade instructions. Regarding format 2 kernel support, all my kernels do (Debian Jessie 3.16 at least) alas I doubt they support the features added in Jewel, so same diff, incompatible default settings when upgrading to it. To give you a real life example, my main Ceph use case is is RBD with a ganeti cluster, each ganeti compute node hosts about 120 VMs. While ganeti actually uses the qemu librbd for actual access it ALSO maps the RBD images via the kernel module because of "reasons". ^o^ So lets assume we need to upgrade the kernel to 4.4 or later on a compute node, this will involve live migrating all VMs to the n+1 spare node. Which can only be done in the weekend maintenance window, at about 20 VMs per window. At best a 1 month procedure. Lastly, your mail client does top-quoting (not going to argue about that) AND puts your signature on top as well, thus causing my mail client to not quote anything below it. Christian -- Christian Balzer Network/Systems Engineer chibi@xxxxxxx Global OnLine Japan/Rakuten Communications http://www.gol.com/ _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |