On 8 July 2016 at 14:34, Ravishankar N <ravishankar at redhat.com> wrote:> There is no chunk size; granular-entry heal is for faster healing of > directory entries, not healing specific portions within a given file. > The non-granular way of doing this was to do an 'expunge+impunge' of the > parent directory.Ok, that explains why the heal was so slow :) So what happened to granular data heals? I thought it was one of the big new features in 3.8. As described here: http://www.gluster.org/pipermail/gluster-devel/2015-September/046773.html - 2) Granular data self-heals -- Lindsay
On 07/08/2016 10:25 AM, Lindsay Mathieson wrote:> So what happened to granular data heals? I thought it was one of the > big new features in 3.8. As described here: > > http://www.gluster.org/pipermail/gluster-devel/2015-September/046773.html > > - 2) Granular data self-healsIt is on the cards (the other items mentioned in that email too) but nothing concrete yet. For VM use cases, sharding anyway emulates granular data heal since the VM image is split into shards and AFR heals only the ones that need data heal (as opposed to healing one big VM image if sharding weren't enabled). -Ravi