Description of problem: I had created a volume and mounted it with the -o acl option. I had immediately setfacl on the mount point setfacl -d --set o::rx /mnt-acl [root@v3136a mnt-acl]# getfacl . # file: . # owner: root # group: root user::rwx group::r-x other::r-x default:user::rwx default:group::r-x default:other::r-x I created a sub-directory and the posix_acl_default was propagated: [root@v3136a mnt-acl]# getfacl test1.d # file: test1.d # owner: root # group: root user::rwx group::r-x other::r-x default:user::rwx default:group::r-x default:other::r-x But whenever I did a vi on a file I would get the following errors without fail: [2016-10-17 20:15:36.187211] I [dict.c:473:dict_get] (-->/usr/lib64/glusterfs/3.7.9/xlator/debug/io-stats.so(io_stats_lookup_cbk+0x167) [0x7f85be851a97] -->/usr/lib64/glusterfs/3.7.9/xlator/system/posix-acl.so(posix_acl_lookup_cbk+0x23a) [0x7f85be63c71a] -->/usr/lib64/libglusterfs.so.0(dict_get+0x63) [0x7f85d18086e3] ) 0-dict: !this || key=system.posix_acl_access [Invalid argument] [2016-10-17 20:15:36.192142] I [dict.c:473:dict_get] (-->/usr/lib64/glusterfs/3.7.9/xlator/debug/io-stats.so(io_stats_lookup_cbk+0x167) [0x7f85be851a97] -->/usr/lib64/glusterfs/3.7.9/xlator/system/posix-acl.so(posix_acl_lookup_cbk+0x28c) [0x7f85be63c76c] -->/usr/lib64/libglusterfs.so.0(dict_get+0x63) [0x7f85d18086e3] ) 0-dict: !this || key=system.posix_acl_default [Invalid argument] I would assumed that such errors would not appear from the admin guide: Effects of a Default ACL The following are the ways in which the permissions of a directory's default ACLs are passed to the files and subdirectories in it: A subdirectory inherits the default ACLs of the parent directory both as its default ACLs and as an access ACLs. A file inherits the default ACLs as its access ACLs. The errors are filling up the client log pretty fast. Version-Release number of selected component (if applicable): RHGS 3.1.3 RHEL 6.8 How reproducible: Easliy Steps to Reproduce: 1. mount with the -o acl mount option 2. setfacl -d --set o::rx /mnt-acl 3. vi any file Actual results: posix_acl_default and posix_acl_access messages will appear in the log Expected results: posix_acl_default and posix_acl_access messages should not appear in the log Additional info:
Yes, you are right, the error is only seen with existing file and with stat-prefetch on. The initial analysis shows that if no explicit ACL is set on the file md-cache will not fetch the extended attribute related to ACL. And thus acl xlator is throwing this error. I will do some more analysis and post the findings here.
As mentioned in comment 5 the problem is seen only when stat-prefetch is on and the cached file does not have any acl set. Following patch is posted in upstream to prevent the unnecessary logs generated. http://review.gluster.org/15769/
Downstream patch: https://code.engineering.redhat.com/gerrit/89419
upstream mainline : http://review.gluster.org/15769 upstream 3.9 : http://review.gluster.org/15801 downstream patch : https://code.engineering.redhat.com/gerrit/#/c/89419/
Upstream 3.8 patch: http://review.gluster.org/15799 Upstream 3.7 patch: http://review.gluster.org/15798
acked+ for the accelerated fix. Please ensure that the fix makes it to 3.2 as well.
I am able to reproduce this issue using 3.1.3 build using the steps mentioned in the description section of bug. Repeated the same steps using the rhgs3.2 build ( glusterfs-3.8.4-5.el6rhs.x86_64 ),i didn't see the issue reported, Fix is working GOOD, This testing is done on RHEL6 Env.
(In reply to Byreddy from comment #18) > I am able to reproduce this issue using 3.1.3 build using the steps > mentioned in the description section of bug. > > Repeated the same steps using the rhgs3.2 build ( > glusterfs-3.8.4-5.el6rhs.x86_64 ),i didn't see the issue reported, Fix is > working GOOD, > > > This testing is done on RHEL6 Env. Tested same thing on RHEL7 as well using glusterfs-3.8.4-5.el7rhgs.x86_64 version bits, it's working fine. I will test with hot fix build provided and update the result here.
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://rhn.redhat.com/errata/RHSA-2017-0486.html