Hi Lindsay, As I understand it, the current logic of bitd/scrubd does not address the problem you asked about "a process where all replicas are compared for inconsistencies."
bitd/scrubd operate independently within each node, signing each file and validating the checksum - which is part of the answer. It basically protects the files on a brick from silent data corruption - it does not ensure that each replica is consistent with each other. I believe that replica compare and automatic healing, are features being worked on for a future release - not sure of timing, but maybe one of the devs can chime in. PC On Thu, Apr 14, 2016 at 5:54 AM, Kaleb S. KEITHLEY <[email protected]> wrote: > On 04/13/2016 01:41 PM, Alan Millar wrote: > >> Please refer to Bitrot Feature: > > > >> > http://www.gluster.org/community/documentation/index.php/Features/BitRot > > > >> I suppose it is already quite mature because it was already listed as a > > > >> feature on RedHat Gluster Storage 3.2 Administration Guide. > > > > If you are only looking for *detection*, it sounds like it is mature and > working. If you want automatic *correction* on a replicated volume (like > you find in ZFS or BTRFS), that doesn't appear to exist yet. > > > > > > I can't find an RHGS 3.2 admin guide, but I can find an RHGS 3.1 guide at > > > > RHGS 3.2 hasn't been released yet. RHGS 3.1.2 is the current version. > > -- > > Kaleb > _______________________________________________ > Gluster-users mailing list > [email protected] > http://www.gluster.org/mailman/listinfo/gluster-users >
_______________________________________________ Gluster-users mailing list [email protected] http://www.gluster.org/mailman/listinfo/gluster-users
