Description of problem: Default ACLs are not synced to geo-replication slave. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. On geo-rep master: mkdir dir1 setfacl -d -m g::rwx dir1 getfacl dir1 2. On geo-rep slave: getfacl dir1 Actual results: Default ACL is not set on slave. Expected results: Default ACL is set on slave. Additional info: Access ACLs are synced as expected. With slave-gluster-log-level=DEBUG, the following log messages appears. [2019-07-17 03:49:00.305670] D [fuse-helpers.c:649:fuse_ignore_xattr_set] 0-glusterfs-fuse: disallowing setxattr: key [system.posix_acl_default], client pid [-1] [2019-07-17 03:49:00.308140] D [fuse-helpers.c:649:fuse_ignore_xattr_set] 0-glusterfs-fuse: disallowing setxattr: key [trusted.SGI_ACL_DEFAULT], client pid [-1] On the other hand, for access ACL, [2019-07-17 06:11:27.892153] D [fuse-helpers.c:649:fuse_ignore_xattr_set] 0-glusterfs-fuse: allowing setxattr: key [system.posix_acl_access], client pid [-1] In fuse_ignore_xattr_set(), only the following xattrs are allowed to be set by gsyncd: /* trusted NS check */ if (!((fnmatch("*.glusterfs.*.xtime", key, FNM_PERIOD) == 0) || (fnmatch("*.glusterfs.volume-mark", key, FNM_PERIOD) == 0) || (fnmatch("*.glusterfs.volume-mark.*", key, FNM_PERIOD) == 0) || (fnmatch("system.posix_acl_access", key, FNM_PERIOD) == 0) || (fnmatch("glusterfs.gfid.newfile", key, FNM_PERIOD) == 0) || (fnmatch("*.glusterfs.shard.block-size", key, FNM_PERIOD) == 0) || (fnmatch("*.glusterfs.shard.file-size", key, FNM_PERIOD) == 0))) ret = -1; out: gf_log("glusterfs-fuse", GF_LOG_DEBUG, "%s setxattr: key [%s], " " client pid [%d]", (ret ? "disallowing" : "allowing"), key, priv->client_pid); Also, on hybrid crawl, only MKDIR and SETATTR changelog records are genereated for directories. I think SETXATTR records should also be generated.
This bug is moved to https://github.com/gluster/glusterfs/issues/959, and will be tracked there from now on. Visit GitHub issues URL for further details