“No data available” Error in GlusterFS Brick Log

Recently I was poring over my GlusterFS brick logs trying to troubleshoot a problem with self healing and I saw some (6,000+!) errors which alarmed me:

[2013-11-22 09:21:51.138732] E [posix.c:2668:posix_getxattr] 0-homes-posix: getxattr failed on /mnt/gfs/wingu0/sda1/homes/yajamma/qdd_programs/qdd_files/qdd3_beta/MS_extract.pl: system.posix_acl_access (No data available)

It turns out this error isn’t an error after all, and there’s a patch merged upstream to re-classify this to DEBUG severity. We’re running GlusterFS 3.4.0 right now, and GlusterFS 3.4.1 was released before this patch was merged, so I guess it will land in 3.4.2(?).

As to why this error is popping up, I’m not sure. I’d be inclined to think my underlying storage doesn’t support ACLs, but we’re using XFS and access control lists are always enabled on XFS. In any case, unless you’re a GlusterFS engineer, you can safely ignore this “error”.

Related posts (automagically generated):

  1. Troubleshooting GlusterFS performance issues
  2. Replacing Failed Drive in Hardware RAID5 on GlusterFS Replica
  3. Fun with kernels on the Huawei U8150
  4. Update GlusterFS 3.3.1 to 3.4.0 on CentOS 6.4 cluster
  5. Building CyanogenMod 7.1 for the Huawei U8150