Description of problem:

 Auxiliary mount remains even after crawler finishes crawling , when quota is enabled when bricks are down.

How reproducible:

1/1

Steps to Reproduce:
1. Kill one of the brick in each sub-volume
2. Enable quota 
3. 

Actual results:

[root@rhs-arch-srv3 tmp]# ps -ef | grep quota
root      7762     1  0 07:01 ?        00:00:00 /usr/sbin/glusterfs --volfile-server localhost --volfile-id clone2 -l /var/log/glusterfs/quota-mount-clone2.log -p /var/run/gluster/clone2.pid --client-pid -5 /var/run/gluster/clone2/
root     14877  6664  0 09:20 pts/0    00:00:00 grep quota
root     15095     1  0 Mar01 ?        00:00:01 /usr/sbin/glusterfs -s localhost --volfile-id client_per_brick/clone2.client.10.70.36.3.var-run-gluster-snaps-0c286cc539c34063aae2d17fdb66fd71-brick2-b2.vol --use-readdirp=yes --client-pid -100 -l /var/log/glusterfs/quota_crawl/var-run-gluster-snaps-0c286cc539c34063aae2d17fdb66fd71-brick2-b2.log /var/run/gluster/tmp/mntRlTPnv
root     18829     1  0 Mar01 ?        00:00:02 /usr/sbin/glusterfs -s localhost --volfile-id gluster/quotad -p /var/lib/glusterd/quotad/run/quotad.pid -l /var/log/glusterfs/quotad.log -S /var/run/gluster/3b3b5f5b2555b948d2dbb20210fbc9d6.socket --xlator-option *replicate*.data-self-heal=off --xlator-option *replicate*.metadata-self-heal=off --xlator-option *replicate*.entry-self-heal=off


glusterd logs:

 [2017-03-01 06:02:36.883481] W [MSGID: 106033] [glusterd-quota.c:331:_glusterd_quota_initiate_fs_crawl] 0-management: chdir /var/run/gluster/tmp/mntRlTPnv failed [Transport endpoint is not connected]


Since the brick is down we get transport_disconnected on chdir to mountpoint.

In the function _glusterd_quota_initiate_fs_crawl, we exit if chdir fails without doing a umount. 

                ret = chdir (mountdir);
                if (ret == -1) {
                        gf_msg (THIS->name, GF_LOG_WARNING, errno,
                                GD_MSG_DIR_OP_FAILED, "chdir %s failed",
                                mountdir);
                        exit (EXIT_FAILURE);
                }

To fix this we need to do a lazy_umount before exit.