Am 01.10.2012 13:29, schrieb Karel Zak: > On Sat, Sep 29, 2012 at 07:36:54PM +0200, Reindl Harald wrote: > There is no difference. The "bind" is an operation, not a special > state of any mountpoint. Nowhere in the system is information that > the mountpoint has been created by "bind" -- the kernel does not > see any difference between the original and bind mount. It's just > another reference to the same object (device). > mount /dev/sda1 /mnt1 > mount /dev/sda1 /mnt2 > > is exactly the same as: > > mount /dev/sda1 /mnt1 > mount --bind /mnt1 /mnt2
but there was a difference until F15 came out with systemd > We should not care about "bind" flag in mtab (or another place) at > all. The solution is to de-duplicate df(1) output. and how should "df" do this? how should "df" smell that "/mnt/data" is the one to display? /dev/md2 3814414416 2335948104 1478466312 62% /mnt/data /dev/md2 3814414416 2335948104 1478466312 62% /home /dev/md2 3814414416 2335948104 1478466312 62% /tmp /dev/md2 3814414416 2335948104 1478466312 62% /var/tmp /dev/md2 3814414416 2335948104 1478466312 62% /Volumes/dune/www-servers /dev/md2 3814414416 2335948104 1478466312 62% /Volumes/dune/www-servers/phpincludes > It's fine to blame systemd guys for all the bad things in the World, > but kill the original mtab concept was planned independently on > systemd. however, before killing things which worked fine for decades there should be a thought how to change things internally without change the behavior from the users view only the fact that the named-chroot mounts caused a error message for a normal user proves that there was not much testing/thinking about impacts - this one meant to get error-mails from EVERY cronjob-script using "df"
signature.asc
Description: OpenPGP digital signature
_______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel