Ensuring that rsync doesn't try to write to an unmounted drive

Chris cpollock at embarqmail.com
Tue Sep 11 18:06:42 UTC 2018

On Tue, 2018-09-11 at 12:56 -0400, Kevin Korb via rsync wrote:
> --timeout is about network connection timeouts.  You aren't using the
> network so it doesn't apply at all.  Even if you were networking an
> unmounted filesystem is an empty directory as far as rsync is
> concerned
> and rsync would treat it that way with no idea that you intended to
> have
> something mounted there.
Thanks Kevin, didn't realize that. Missed that when reading about --

> Now, I see at the top of your script you check for the existence of
> the
> target directory.  If that isn't the root of a filesystem then you
> are
> good because you are already checking for that.  If it is the root of
> the filesystem then it will exist either as an empty directory or a
> mount point and you need to check for those possibilities.

I hope I'm answering your question. The mount info on the drive is

/dev/sdb1 on /media/chris/backup2 type ext4

However, in the script it's actually checking for the folder the backup
is written to 'snapshot'

if [ ! -d ${BACKUP_DIR} ];then
    echo "Backup destination directory ${BACKUP_DIR} not exist."
        echo "run 'sudo mkdir ${BACKUP_DIR}' to create. "
    exit 1 

So if I understand the script is checking for the 'snapshot' folder on
the mount point 'backup2'. Is that correct? If rsync can't see the
'snapshot' folder would it still turn around and write or attempt to
write the backup to the hard drive or would it gracefully fail?

I had asked a question on LQ.org about the script earlier because I
couldn't see why it wasn't making a log file. Note-this script was
written by someone here on the list for me about 4 or so years ago.
Yes, it took me that long to notice. One of the replies to my question
suggested that I add this to the script:


df -h | grep $mount_point > /dev/null
if [ $? -eq 0 ]

echo "mount point $mount_point exists, rsync started"
    echo "Error: mount point $mount_point does not exist, rsync
operation skipped"

So, would it be ok to leave it as is or do I need to add the above to
check for the actual mount point?

> On 09/11/2018 12:28 PM, Chris via rsync wrote:
> > I have a script that runs nightly as a cronjob to backup my drive
> > to a
> > USB drive https://pastebin.com/yivqrGUC On the command line I use
> > the
> > --timeout option. Is this sufficient to ensure that if the external
> > drive somehow becomes unmounted that rsync will gracefully fail
> > without
> > trying to write to the hard drive instead of the USB drive?
> > 
> > rsync -vaWSHpl --timeout=15 --delete-excluded --filter "merge
> > ${EXC_FILE}" / "${BACKUP_DIR}" > /home/chris/rsyncbackup.log 2>
> > /home/chris/rsyncbackup.errors.log 
> > 
> > If this is sufficient or would it be better if I lowered the
> > 'timeout'
> > to 5 seconds?
> > 
> > Thanks for any suggestions/advice.
> > 
> > Chris
> > 
> > 
> > 
KeyID 0xE372A7DA98E6705C
31.11972; -97.90167 (Elev. 1092 ft)
12:50:18 up 19:57, 1 user, load average: 0.82, 0.92, 1.29
Description:	Ubuntu 18.04.1 LTS, kernel 4.15.0-34-generic
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: This is a digitally signed message part
URL: <http://lists.samba.org/pipermail/rsync/attachments/20180911/0fad40d5/signature.sig>

More information about the rsync mailing list