[Gluster-users] dir delete operation error Transport endpoint is not connected

[Satish Patel]

Folk,

I am dealing with a very strange issue related to Glusterfs and not
sure if its volume issue or bug. I have been provided mount point for
my project which i have mounted on my centos 8 server using following
command

mount -t glusterfs x.x.x.x:gluster_vol2/voyager /root/m/

Now I can create files and delete file without issue but at the same
time if I create a directory it works but i can't able to delete the
directory. I get the following error when I delete a directory.

$ mkdir testdir1
$ mkdir testdir1/testdir
$ rmdir testdir1/testdir
rmdir: failed to remove 'testdir': Transport endpoint is not connected

$[ rm -rf testdir1
rm: cannot remove 'testdir1/testdir': Transport endpoint is not connected

I have no control on the glusterfs server side, I asked the admin and
they said everything looks ok on their side. I am seeing following
error on my client side logs

[2022-02-08 03:18:52.451894 +0000] W [dict.c:1532:dict_get_with_ref]
(-->/usr/lib64/glusterfs/9.5/xlator/cluster/distribute.so(+0x51586)
[0x7f2451626586] -->/lib64/libglusterfs.so.0(dict_get_uint32+0x3b)
[0x7f245ab1457b] -->/lib64/libglusterfs.so.0(dict_get_with_ref+0x85)
[0x7f245ab13975] ) 0-dict: dict OR key
(trusted.glusterfs.dht.commithash) is NULL [Invalid argument]
[2022-02-08 03:18:53.780262 +0000] W [dict.c:1532:dict_get_with_ref]
(-->/usr/lib64/glusterfs/9.5/xlator/cluster/distribute.so(+0x51586)
[0x7f2451626586] -->/lib64/libglusterfs.so.0(dict_get_uint32+0x3b)
[0x7f245ab1457b] -->/lib64/libglusterfs.so.0(dict_get_with_ref+0x85)
[0x7f245ab13975] ) 0-dict: dict OR key
(trusted.glusterfs.dht.commithash) is NULL [Invalid argument]
[2022-02-08 03:18:53.783392 +0000] W
[fuse-bridge.c:2052:fuse_unlink_cbk] 0-glusterfs-fuse: 482: RMDIR()
/testdir1/testdir => -1 (Transport endpoint is not connected)