Description of problem:
======================
Brick goes down in the middle of this particular test case

On the Master:
--------------
[2018-10-18 09:30:58.93183] I [master(/rhs/brick2/b5):1450:crawl] _GMaster: slave's time        stime=(1539855011, 0)
[2018-10-18 09:31:00.624412] E [repce(/rhs/brick3/b8):209:__call__] RepceClient: call failed    call=17069:139754185205568:1539855057.49        method=entry_ops        error=OSError
[2018-10-18 09:31:00.625464] E [syncdutils(/rhs/brick3/b8):349:log_raise_exception] <top>: Gluster Mount process exited error=ENOTCONN
[2018-10-18 09:31:00.699959] I [syncdutils(/rhs/brick3/b8):295:finalize] <top>: exiting.


brick3/b8 logs report:
----------------------
[2018-10-18 09:31:00.725836] W [socket.c:593:__socket_rwv] 0-master-changelog: readv on /var/run/gluster/.f8271615d91fca5417068.sock failed (No data available)
[2018-10-18 09:31:00.739737] I [MSGID: 115036] [server.c:571:server_rpc_notify] 0-master-server: disconnecting connection from dhcp42-2.lab.eng.blr.redhat.com-17140-2018/10/18-09:30:01:598538-master-client-4-0-0
[2018-10-18 09:31:00.740127] I [MSGID: 101055] [client_t.c:443:gf_client_unref] 0-master-server: Shutting down connection dhcp42-2.lab.eng.blr.redhat.com-17140-2018/10/18-09:30:01:598538-master-client-4-0-0


On the slave:
-------------
Traceback (most recent call last):
  File "/usr/libexec/glusterfs/python/syncdaemon/repce.py", line 113, in worker
    res = getattr(self.obj, rmeth)(*in_data[2:])
  File "/usr/libexec/glusterfs/python/syncdaemon/resource.py", line 639, in entry_ops
    st = lstat(slink)
  File "/usr/libexec/glusterfs/python/syncdaemon/syncdutils.py", line 577, in lstat
    return errno_wrap(os.lstat, [e], [ENOENT], [ESTALE, EBUSY])
  File "/usr/libexec/glusterfs/python/syncdaemon/syncdutils.py", line 559, in errno_wrap
    return call(*arg)
OSError: [Errno 107] Transport endpoint is not connected: '.gfid/4b67b1d8-b53b-4962-a4c4-294b3d5e750c'



Version-Release number of selected component (if applicable):
=============================================================
[root@dhcp42-2 bricks]# rpm -qa | grep gluster
glusterfs-server-3.12.2-22.el7rhgs.x86_64
glusterfs-client-xlators-3.12.2-22.el7rhgs.x86_64
glusterfs-rdma-3.12.2-22.el7rhgs.x86_64
vdsm-gluster-4.19.43-2.3.el7rhgs.noarch
gluster-nagios-common-0.2.4-1.el7rhgs.noarch
glusterfs-3.12.2-22.el7rhgs.x86_64
glusterfs-api-3.12.2-22.el7rhgs.x86_64
glusterfs-events-3.12.2-22.el7rhgs.x86_64
glusterfs-libs-3.12.2-22.el7rhgs.x86_64
glusterfs-fuse-3.12.2-22.el7rhgs.x86_64
glusterfs-geo-replication-3.12.2-22.el7rhgs.x86_64
python2-gluster-3.12.2-22.el7rhgs.x86_64
gluster-nagios-addons-0.2.10-2.el7rhgs.x86_64
glusterfs-cli-3.12.2-22.el7rhgs.x86_64
libvirt-daemon-driver-storage-gluster-4.5.0-10.el7.x86_64
[root@dhcp42-2 bricks]# 


How reproducible:
=================
2/2

Steps to Reproduce:
===================
1.Create and start a master and slave arbiter volume
2.Set up a geo-rep session between the 2
3.Mount the master and pump IO:

for i in {create,chmod,symlink,create,chown,chmod,create,symlink,chgrp,symlink,truncate,symlink,chown,create,symlink}; do crefi --multi -n 5 -b 10 -d 10 --max=10K --min=500 --random -T 10 -t text --fop=$i /mnt/master/ ; sleep 10 ; done


master vol info:
-----------------
Volume Name: master
Type: Distributed-Replicate
Volume ID: 5a97408a-8cc0-4f24-a306-7f9e143e6614
Status: Started
Snapshot Count: 0
Number of Bricks: 2 x (2 + 1) = 6
Transport-type: tcp
Bricks:
Brick1: 10.70.43.116:/rhs/brick2/b4
Brick2: 10.70.42.2:/rhs/brick2/b5
Brick3: 10.70.42.44:/rhs/brick2/b6 (arbiter)
Brick4: 10.70.43.116:/rhs/brick3/b7
Brick5: 10.70.42.2:/rhs/brick3/b8
Brick6: 10.70.42.44:/rhs/brick3/b9 (arbiter)
Options Reconfigured:
changelog.changelog: on
geo-replication.ignore-pid-check: on
geo-replication.indexing: on
transport.address-family: inet
nfs.disable: on
performance.client-io-threads: off
cluster.brick-multiplex: off
cluster.enable-shared-storage: enable


slave vol info:
---------------
Volume Name: slave
Type: Distributed-Replicate
Volume ID: 125aece0-1800-4065-a363-f36dc0efc6f5
Status: Started
Snapshot Count: 0
Number of Bricks: 2 x (2 + 1) = 6
Transport-type: tcp
Bricks:
Brick1: 10.70.42.226:/rhs/brick2/b4
Brick2: 10.70.43.81:/rhs/brick2/b5
Brick3: 10.70.41.204:/rhs/brick2/b6 (arbiter)
Brick4: 10.70.42.226:/rhs/brick3/b7
Brick5: 10.70.43.81:/rhs/brick3/b8
Brick6: 10.70.41.204:/rhs/brick3/b9 (arbiter)
Options Reconfigured:
transport.address-family: inet
nfs.disable: on
performance.client-io-threads: off



Actual results:
==============
Brick went down

Expected results:
================
Brick should not go down