[Gluster-devel] Incorrect exit code from mount.glusterfs

[Niels de Vos]

On Tue, Apr 08, 2014 at 03:50:42PM +0530, Deepak Shetty wrote:
> Hi All,
>    I am wondering if this I am the only one seeing this or there are enuf
> reasons why mount.glusterfs returns 0 (which means success) as the exit
> code for error cases ?
> Bcos of this cinder (openstack service) code is misled as it thinks
> mounting glusterfs volume on already mounted volume is successfull and
> never gets into the warning/error flow!
> (Not to mention that i spent 1+ days debugging and reaching that
> conclusion!!!)

This should definitely get fixed in the mount.glusterfs shell script.  
The correct return values are listed in 'man 8 mount'. Please file a bug 
for this:
- https://bugzilla.redhat.com/enter_bug.cgi?product=GlusterFS&component=scripts

Thanks,
Niels

> 
> I just did a quick sanity check to compare how mount.nfs and
> mount.glusterfs behave on similar error scenario.. and below is what i find
> 
>  [stack at devstack-vm cinder]$ df -h
> Filesystem Size Used Avail Use% Mounted on
> /dev/vda1 9.9G 3.7G 6.1G 38% /
> devtmpfs 2.0G 0 2.0G 0% /dev
> tmpfs 2.0G 0 2.0G 0% /dev/shm
> tmpfs 2.0G 448K 2.0G 1% /run
> tmpfs 2.0G 0 2.0G 0% /sys/fs/cgroup
> 192.168.122.252:/opt/stack/nfs/brick 9.9G 3.7G 6.1G 38%
> /opt/stack/data/cinder/mnt/f23011fcca5ae3a8b8ebfd7e4af2e190
> 
> [stack at devstack-vm cinder]$ sudo mount -t nfs
> 192.168.122.252:/opt/stack/nfs/brick
> /opt/stack/data/cinder/mnt/f23011fcca5ae3a8b8ebfd7e4af2e190/
> mount.nfs: /opt/stack/data/cinder/mnt/f23011fcca5ae3a8b8ebfd7e4af2e190 is
> busy or already mounted
> 
> *[stack at devstack-vm cinder]$ echo $? 32*
> NOTE: mount.nfs exits w/ proper error code
> 
>  [stack at devstack-vm ~]$ df -h
> Filesystem Size Used Avail Use% Mounted on
> /dev/vda1 9.9G 3.7G 6.1G 38% /
> devtmpfs 2.0G 0 2.0G 0% /dev
> tmpfs 2.0G 0 2.0G 0% /dev/shm
> tmpfs 2.0G 448K 2.0G 1% /run
> tmpfs 2.0G 0 2.0G 0% /sys/fs/cgroup
> devstack-vm.localdomain:/gvol1 9.9G 3.7G 6.1G 38%
> /opt/stack/data/cinder/mnt/d45ccec4f1572f6f242b70befa3d80fe
> devstack-vm.localdomain:/gvol2 9.9G 3.7G 6.1G 38%
> /opt/stack/data/cinder/mnt/413c1f8d14058d5b2d07f8a92814bd12
> 
> [stack at devstack-vm ~]$ sudo mount -t glusterfs
> devstack-vm.localdomain:/gvol1
> /opt/stack/data/cinder/mnt/d45ccec4f1572f6f242b70befa3d80fe/
> /sbin/mount.glusterfs: according to mtab, GlusterFS is already mounted on
> /opt/stack/data/cinder/mnt/d45ccec4f1572f6f242b70befa3d80fe
> 
> *[stack at devstack-vm ~]$ echo $? 0*
>  NOTE: mount.glusterfs exits with 0 (success)
> 
> ******************************************************************************************
> 
> A quick look at mount.glusterfs yeilds...
> 
>     # No need to do a ! -d test, it is taken care while initializing the
>     # variable mount_point
>     [ -z "$mount_point" -o ! -d "$mount_point" ] && {
>         echo "ERROR: Mount point does not exist."
>         usage;
> *        exit 0;*
>     }
> 
>     # Simple check to avoid multiple identical mounts
>     if grep -q "[[:space:]+]${mount_point}[[:space:]+]fuse" $mounttab; then
>         echo -n "$0: according to mtab, GlusterFS is already mounted on "
>         echo "$mount_point"
> *        exit 0;*
>     fi
> 
> ******************************************************************
> 
> Is this intended or bug or is there some history to why mount.glusterfs
> return 0 for many obvious error cases ?
> 
> thanx,
> deepak

> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at nongnu.org
> https://lists.nongnu.org/mailman/listinfo/gluster-devel