FS#70693 - [pam_mount] FUSE & server (e.g. NFS) mounts not detected as already existing
Attached to Project:
Community Packages
Opened by Oliver Ford (OJFord) - Monday, 03 May 2021, 16:20 GMT
Last edited by Buggy McBugFace (bugbot) - Saturday, 25 November 2023, 20:02 GMT
Opened by Oliver Ford (OJFord) - Monday, 03 May 2021, 16:20 GMT
Last edited by Buggy McBugFace (bugbot) - Saturday, 25 November 2023, 20:02 GMT
|
Details
Description:
pam_mount will fail to detect that a mount already exists, and attempt to re-mount it in another session, if using either `fstype="fuse"` or a non-null `server`. Additional info: * v2.18 * upstream bug report: https://sourceforge.net/p/pam-mount/bugs/117/ * Debian bug report: https://bugs.launchpad.net/debian/+source/libpam-mount/+bug/1908638 * patch for the `server` case *only*: https://sourceforge.net/p/pam-mount/pam-mount/merge-requests/3/ Steps to reproduce: 1. Configure pam_mount.conf.xml with a FUSE filesystem, e.g. gocryptfs: ``` <volume user="%(USER)" fstype="fuse" options="nodev,nosuid,quiet,nonempty,allow_other" path="/usr/bin/gocryptfs#/home/%(USER).cipher" mountpoint="/home/%(USER)" /> ``` 2. Log in on tty1, observe it mounted (`mount -l`) 3. Log in on tty2, observe it re-mounted (because of allowing `nonempty` I believe, the behaviour on re-mount attempt may vary with different filesystems and the mount configuration) 4. Log out of tty2, observe 'transport endpoint is not connected' on tty1, no home dir, two mounts still in `mount -l` 5. `fusermount -uz $HOME` to rectify |
This task depends upon
Closed by Buggy McBugFace (bugbot)
Saturday, 25 November 2023, 20:02 GMT
Reason for closing: Moved
Additional comments about closing: https://gitlab.archlinux.org/archlinux/p ackaging/packages/pam_mount/issues/1
Saturday, 25 November 2023, 20:02 GMT
Reason for closing: Moved
Additional comments about closing: https://gitlab.archlinux.org/archlinux/p ackaging/packages/pam_mount/issues/1
https://git.launchpad.net/ubuntu/+source/libpam-mount/tree/debian/patches/0014-Don-t-compare-source-when-checking-if-it-s-already-m.patch
which addresses the `fstype="fuse"` use case, among others by the look of its commit message, but is the source of the problem for (some? all?) NFS use cases.