[Gluster-users] Stale File handle
Hello, I have a file in my gluster volume that has the following when i try to access it (ls) ls: cannot access 37019600-c34e-4d10-8829-ac08cb141f19.meta: Stale file handle 37019600-c34e-4d10-8829-ac08cb141f19 37019600-c34e-4d10-8829-ac08cb141f19.lease 37019600-c34e-4d10-8829-ac08cb141f19.meta when i look at gluster volume heal VMData info i get the following Brick found1.ssd.org:/data/brick1/data /8d4b29ee-16c9-4bb9-a7db-937c9da6805d/images/a01b2883-73b1-4d4f-a338-44afcfce57a6/37019600-c34e-4d10-8829-ac08cb141f19 Status: Connected Number of entries: 1 Brick found2.ssd.org:/data/brick1/data Status: Connected Number of entries: 0 Brick found3.ssd.org:/data/brick1/arbit /8d4b29ee-16c9-4bb9-a7db-937c9da6805d/images/a01b2883-73b1-4d4f-a338-44afcfce57a6/37019600-c34e-4d10-8829-ac08cb141f19 Status: Connected Number of entries: 1 the entries never sync! if i stop the volume and examine the file on the brick itself the file is fine but i cannot get gluster to accept the file on disk as authoritative. any ideas how to fix this Using gluster 3.12.6 Thanks Russell Wecker IT Director Southern Asia-Pacific Division San Miguel II Bypass Silang, Silang, Cavite 4118 Philippines Phone: +63 46 414 4000 x 5210 Cell: +63 917 595 6395 URL: [ http://ssd.adventist.asia/ | http://ssd.adventist.asia ] ___ Gluster-users mailing list Gluster-users@gluster.org http://lists.gluster.org/mailman/listinfo/gluster-users
Re: [Gluster-users] Failover problems with gluster 3.8.8-1 (latest Debian stable)
On Fri, Feb 16, 2018 at 05:44:43AM -0600, Dave Sherohman wrote: > On Thu, Feb 15, 2018 at 09:34:02PM +0200, Alex K wrote: > > Have you checked for any file system errors on the brick mount point? > > I hadn't. fsck reports no errors. > > > What about the heal? Does it report any pending heals? > > There are now. It looks like taking the brick offline to fsck it was > enough to trigger gluster to recheck everything. I'll check after it > finishes to see whether this ultimately resolves the issue. Finally got a chance to test it and, nope, still having problems. Although I'm sure the gluster volume is in better condition now than before I got this housekeeping to run, it doesn't seem to have actually changed the final results of azathoth going down at all. Anyone have a more thorough way of forcing a replica to refresh itself from its twin? -- Dave Sherohman ___ Gluster-users mailing list Gluster-users@gluster.org http://lists.gluster.org/mailman/listinfo/gluster-users