Description of problem:
Customer has run stop-all-gluster-processes.sh as part of their preparation for upgrade. During the run this for:

From line 75 from https://github.com/gluster/glusterfs/blob/master/extras/stop-all-gluster-processes.sh#L75 :
```
for pidfile in $(find /var/run/gluster/ -name '*.pid');
do
    local pid=$(cat ${pidfile});
    echo "sending SIG${signal} to pid: ${pid}";
    kill -${signal} ${pid};
done
```

found pid files stored in /var/run/gluster/snaps/ from mysql container volume, which resulted in the loop trying to kill pid 1.

root@hostname:/usr/share/glusterfs/scripts # ./stop-all-gluster-processes.sh -g
sending SIGTERM to pid: 1288
sending SIGTERM to pid: 25556
sending SIGTERM to pid: 25275
sending SIGTERM to pid: 22927
sending SIGTERM to pid: 22464
sending SIGTERM to pid: 22243
sending SIGTERM to pid: 8441
sending SIGTERM to pid: 8521
sending SIGTERM to pid: 8354
sending SIGTERM to pid: 8317
sending SIGTERM to pid: 8318
sending SIGTERM to pid: 8319
sending SIGTERM to pid: 1449
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1
sending SIGTERM to pid: 1

sending SIGKILL to pid: 8521
./stop-all-gluster-processes.sh: line 79: kill: (8521) - No such process
sending SIGKILL to pid: 1449
./stop-all-gluster-processes.sh: line 79: kill: (1449) - No such process
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1
sending SIGKILL to pid: 1

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

Red Hat Enterprise Linux Server release 7.5 (Maipo)
redhat-storage-server-3.3.1.0-1.el7rhgs.noarch
glusterfs-server-3.8.4-54.8.el7rhgs.x86_64

How reproducible:
Every time Customer runs the script for maintenance

Actual results:
The script finds all pid processes even from unmounted snaps, this can lead to random pids being killed.

Expected results:
Script only picks up pids for gluster processes, not random pids from /var/run/gluster/snaps.

Additional info: