Description of problem: I have glusterFS with three bricks. Two of them on Ubuntu 12.04 and one on Ubuntu 14.04. When I am trying to write files from the node running Ubuntu 14.04, the following errors/warnings occur (see logs below). However the files are successfully written and they are accessible from all the the nodes. If I try to write the files directly to the folder used as brick, the error does not occur. It only occurs when writing thought the mount point. I am mounting the volume with {THIS_NODE}:nova-gluster-vol /var/lib/nova/instances/ glusterfs defaults,nobootwait ---> Log From Client ( /var/log/glusterfs/{VOLUME}) [2015-07-23 08:29:07.733023] E [iobuf.c:733:iobuf_unref] (-->/usr/lib/x86_64-linux-gnu/libglusterfs.so.0(call_resume+0x5cf) [0x7f040265510f] (-->/usr/lib/x86_64-linux-gnu/libglusterfs.so.0(default_readv_resume+0x14e) [0x7f040263bcbe] (-->/usr/lib/x86_64-linux-gnu/glusterfs/3.5.4/xlator/performance/quick-read.so(qr_readv+0x4e) [0x7f03fced721e]))) 0-iobuf: invalid argument: iobuf [2015-07-23 08:29:07.733111] E [iobuf.c:839:iobref_unref] (-->/usr/lib/x86_64-linux-gnu/libglusterfs.so.0(call_resume+0x5cf) [0x7f040265510f] (-->/usr/lib/x86_64-linux-gnu/libglusterfs.so.0(default_readv_resume+0x14e) [0x7f040263bcbe] (-->/usr/lib/x86_64-linux-gnu/glusterfs/3.5.4/xlator/performance/quick-read.so(qr_readv+0x4e) [0x7f03fced721e]))) 0-iobuf: invalid argument: iobref [2015-07-23 08:29:07.870181] W [fuse-bridge.c:1236:fuse_err_cbk] 0-glusterfs-fuse: 1779227: REMOVEXATTR() /_base/7f5ed776d0594ba0d82a16c8de31f43bc18f2ddf => -1 (No data available) [2015-07-23 08:29:07.880319] W [fuse-bridge.c:1236:fuse_err_cbk] 0-glusterfs-fuse: 1779229: REMOVEXATTR() /_base/7f5ed776d0594ba0d82a16c8de31f43bc18f2ddf => -1 (No data available) [2015-07-23 08:29:07.928179] W [fuse-bridge.c:1236:fuse_err_cbk] 0-glusterfs-fuse: 1779233: REMOVEXATTR() /_base/c527143e93073b5eaca89e2d8f356e043513aa73 => -1 (No data available) ---> Log From Server - node that brick resides ( /var/log/glusterfs/bricks/{VOLUME}) [2015-07-23 08:07:53.313335] E [marker.c:2573:marker_removexattr_cbk] 0-nova-gluster-vol-marker: No data available occurred while creating symlinks [2015-07-23 08:11:37.497083] E [marker.c:2573:marker_removexattr_cbk] 0-nova-gluster-vol-marker: No data available occurred while creating symlinks [2015-07-23 08:11:45.435734] E [marker.c:2573:marker_removexattr_cbk] 0-nova-gluster-vol-marker: No data available occurred while creating symlinks [2015-07-23 08:11:46.722702] E [marker.c:2573:marker_removexattr_cbk] 0-nova-gluster-vol-marker: No data available occurred while creating symlinks [2015-07-23 08:11:46.723805] E [marker.c:2573:marker_removexattr_cbk] 0-nova-gluster-vol-marker: No data available occurred while creating symlinks Version-Release number of selected component (if applicable): 3.5.4 installed from ppa:gluster/glusterfs-3.5 How reproducible: Steps to Reproduce: 1. Configure a glusterFS with 2 or more bricks 2. Write a file to volume, though mount point 3. Actual results: Expected results: Additional info: This bug or a similar one is fixed at 3.6.x according to https://bugzilla.redhat.com/show_bug.cgi?id=1188064. Is a backport to 3.5.X possible?
Maybe this helps: http://thedumontster.com/post/93787042025/kill-process-or-sacrifice-child
*** This bug has been marked as a duplicate of bug 1192832 ***