Description of problem:
During our attempt to upgrade from GlusterFS 3.7.6 to 3.10.1, we discovered a file corruption issue when using the latest fuse clients (3.10.1 and 3.11.1).
The upgrade of the server backend and clients went fine.  However, while performing product installations onto a Gluster Replica volume using the newer versions of GlusterFS fuse client, fail over half the time due to file corruption during the installation process.  This issue was not present while using the GlusterFS fuse 3.7.6 client.  Word arounds for this issue are either
  a.) Downgrade the GlusterFS client to 3.7.6 (keeping a 3.10.1 or 3.11.1 gluster server backend)
  b.) mounting the Gluster volume as NFS
  c.) inserting slowness in the installation process (turning on debugging) 

Version-Release number of selected component (if applicable):

GlusterFS Version 3.10.1 or 3.11.1 (both server backend and fuse client)
Both latest stable versions show the same symptoms.

How reproducible:

Perform a the creation of a weblogic domain on a gluster replica volume.  Enbedded LDAP Failures occur during admin server startup.  Error messages site embedded LDAP file corruption.  Earlier Versions of Gluster Fuse client, in the same configuration, do not have this issue.

Steps to Reproduce:
1. Setup a GlusterFS cluster of two gluster servers serving up a single replica volume.  This volume is then shared to a number of gluster clients
2. Install WebLogic Applications on the Shared Storage
3. Create a WebLogic Domain and attempt to startup the WebLogic Admin Services

Actual results:
Admin Server Startup Fails.  Error Log sites embedded LDAP credentials are invalid and/or embedded LDAP file corruption. 


Expected results:
WebLogic applications start and function normally (As in the past)


Additional info:
This file corruption issue is only the first finding here.  We expect there may be other scenarios with other similar products.  Will be adding more scenarios once we fully vet this upgrade through our development and test cycle