Forwarding this request from debian to get your thoughts.
--- Begin Message ---
- To: submit@xxxxxxxxxxxxxxx
- Subject: Bug#747948: mount: should warn user when potentially incorrect fstab order
- From: David Guyot <david.guyot@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Tue, 13 May 2014 12:04:50 +0200
- Delivered-to: psusi@localhost
- Delivered-to: submit@xxxxxxxxxxxxxxx
- Delivered-to: util-linux@xxxxxxxxxxxxxxxxxxxxxx
- Reply-to: David Guyot <david.guyot@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>, 747948@xxxxxxxxxxxxxxx
- Resent-cc: LaMont Jones <lamont@xxxxxxxxxx>
- Resent-date: Tue, 13 May 2014 10:06:20 +0000
- Resent-from: David Guyot <david.guyot@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
- Resent-message-id: <handler.747948.B.139997551226595@xxxxxxxxxxxxxxx>
- Resent-sender: Debian QA <qa@xxxxxxxxxxxxxxxxx>
- Resent-to: debian-bugs-dist@xxxxxxxxxxxxxxxx
- User-agent: Mutt/1.5.21 (2010-09-15)
Package: mount Version: 2.20.1-5.3 Severity: wishlist Hello, there. Firstly, tl;dr: mount should warn when interlocked mountpoints in fstab aren't in the expected order. I recently noticed what at first seemed a bug in mount management of fstab, but, according to the french mailing list users, it's more a lack of warning. Here is the problem: currently, at start, filesystems referenced in fstab are mounted following their order in fstab. The problem is, when /etc/fstab contains config for two filesystems with interlocked mountpoints, in my case /var/www and /var/www/cache, they are not necessarily in the right order. In my example, because of limitations in my hosting provider installer and my drive configuration, /var/www/cache configuration was stored before /var/www's one. Because man mount did not contained any warning regarding mountpoints order, I assumed this was automagically managed at start; of course, it proved me wrong, as you can see in the following command dump taken after a reboot: david@Curunir:~$ df -h Sys. fich. Taille Util. Dispo Uti% Monté sur rootfs 54G 3,1G 49G 6% / udev 10M 0 10M 0% /dev tmpfs 13G 328K 13G 1% /run /dev/md1 54G 3,1G 49G 6% / tmpfs 5,0M 0 5,0M 0% /run/lock tmpfs 35G 0 35G 0% /dev/shm /dev/md3 20G 233M 19G 2% /var/log /dev/md4 1,7T 852M 1,6T 1% /var/www/cache /dev/md6 99G 188M 94G 1% /home /dev/md7 1,7T 852M 1,6T 1% /var/www tmpfs 32G 0 32G 0% /tmp david@Curunir:~$ sudo su [sudo] password for david: root@Curunir:/home/david# umount /dev/md4 umount: /var/www/cache: not mounted root@Curunir:/home/david# mount /dev/md4 root@Curunir:/home/david# df -h Sys. fich. Taille Util. Dispo Uti% Monté sur rootfs 54G 3,1G 49G 6% / udev 10M 0 10M 0% /dev tmpfs 13G 328K 13G 1% /run /dev/md1 54G 3,1G 49G 6% / tmpfs 5,0M 0 5,0M 0% /run/lock tmpfs 35G 0 35G 0% /dev/shm /dev/md3 20G 233M 19G 2% /var/log /dev/md4 124G 188M 118G 1% /var/www/cache /dev/md6 99G 188M 94G 1% /home /dev/md7 1,7T 852M 1,6T 1% /var/www tmpfs 32G 4,0K 32G 1% /tmp /dev/md4 124G 188M 118G 1% /var/www/cache As you can see, in this case, the mounting of /var/www/cache behaves erratically when its configuration is placed before /var/www's one in /etc/fstab. When I noticed this problem, I checked all system logs, but none showed messages regarding the mountpoints order, so I suspected a failed disk, but my hosting provider's checks did not show any problem, no more than fsck -f. I almost reinstalled the system, but, at the last moment, I thought of fstab and changed mountpoints order to place /var/www before /var/www/cache, and, after a reboot, the problem was solved. I mailed yesterday the Debian french users mailing list about this problem, which I thought of as a bug, but they explained me that, under some conditions, this behaviour may prove helpful. Nevertheless, I still think that mount should warn about this most likely unwanted behaviour, both in its manpage and by logging warnings in dmesg or syslog, in order to help users confronted to this problem. At first, I expected mount to reject such a mountpoint order to avoid problems, but, if I'm right, it should be a bit extreme and bother users who uses this behaviour to their advantage. So, I would like to request the addition in mount of a warning when it tries to mount a filesystem whose mountpoint contains an already used mountpoint in order to avoid confusion for filesystem administation. This warning should be printed on stderr, if not in dmesg and syslog, to allow the admin to correct mountpoints order in the most likely case the current behaviour is not the expected one. Thanks for reading. Regards. -- David Guyot Administrateur système, réseau et télécommunications / Sysadmin Europe Camions Interactive / Stockway Moulin Collot F-88500 Ambacourt Tel: +33 (0)3 29 30 47 85 Fax : +33 (0)3 29 31 31 31Attachment: signature.asc
Description: Digital signature
--- End Message ---
Attachment:
signature.asc
Description: OpenPGP digital signature
