Description of problem: ======================= 'custom extended attributes' set on a directory are not healed after bringing back the down sub-volumes. Client: ======= getfattr -n user.foo c # file: c user.foo="bar1" Backend bricks: =============== [root@dhcpnode1 distrepx3-b0]# getfattr -d -e hex -m . /bricks/brick0/distrepx3-b0/c getfattr: Removing leading '/' from absolute path names # file: bricks/brick0/distrepx3-b0/c security.selinux=0x73797374656d5f753a6f626a6563745f723a676c7573746572645f627269636b5f743a733000 trusted.gfid=0x326aa2338e3846afa3006e873b313416 trusted.glusterfs.dht=0x00000000000000007ffffffc9ffffffa [root@dhcpnode1 distrepx3-b0]# getfattr -d -e hex -m . /bricks/brick1/distrepx3-b1/c getfattr: Removing leading '/' from absolute path names # file: bricks/brick1/distrepx3-b1/c security.selinux=0x73797374656d5f753a6f626a6563745f723a676c7573746572645f627269636b5f743a733000 trusted.gfid=0x326aa2338e3846afa3006e873b313416 trusted.glusterfs.dht=0x00000000000000009ffffffbbffffff9 Version-Release number of selected component (if applicable): 3.12.2-11.el7rhgs.x86_64 How reproducible: 1/1 Steps to Reproduce: ==================== 1) Create a distributed-replicated volume and start it. 2) FUSE mount it on a client. 3) From client create few directories of depth 3. 4) Now, bring down few dht sub-vols using gf_attach command. (I brought down 2 dht sub-vol and 1 brick in another replica pair) 5) Make metadata changes to the directories. like uid, gid, perms, setxattr 6) Bring back the down sub-vols. 7) Check all the bricks for consistency. Actual results: ================ 'custom extended attributes' are not healed after bringing back the down sub-volumes Expected results: ================= No inconsistencies.
Assigning to Karthik as he's taking a look at this.
Hi, Patch is posted on upstream https://review.gluster.org/#/c/20102/ Regards Mohit Agrawal
Verified this BZ on glusterfs version 3.12.2-14.el7rhgs.x86_64. Followed the same steps mentioned in the description. 'custom extended attributes' set on a directory are healed after bringing back the down sub-volumes. Moving this BZ to Verified.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2018:2607