Bug 6284

Summary:	Boot fails when there are webdav mounts in fstab
Product:	Mageia	Reporter:	Florian Hubold <doktor5000>
Component:	RPM Packages	Assignee:	Colin Guthrie <mageia>
Status:	RESOLVED WORKSFORME	QA Contact:
Severity:	normal
Priority:	Normal	CC:	doktor5000, ueberall
Version:	2
Target Milestone:	---
Hardware:	All
OS:	Linux
Whiteboard:
Source RPM:		CVE:
Status comment:

Description Florian Hubold 2012-06-02 09:27:23 CEST

Boot stalls if there are webdav mounts in fstab, with the following messages:

 Dependency failed. Aborted start of Relabel all filesystems, if necessary [ABORT]
 Dependency failed. Aborted start of Mark the need to relabel after reboot [ABORT]
 Dependency failed. Aborted start of Setup links in /boot for running kernel [ABORT]
 Welcome to emergency mode. Use "systemctl default" or ^D to enter default mode.
 Give root password for maintenance (or type Control-D to continue):

Was also put into Errata: https://wiki.mageia.org/en/Mageia_2_Errata#Mageia_boot_fails_when_there_are_webdav_entries_in_fstab

Please investigate and fix.

Comment 1 Markus Ueberall 2012-06-02 10:26:04 CEST

Boot also stalls and the system goes into emergency mode if the system is unable to mount hibernated Windows partitions. (The fix in both cases should be to either keep retrying in the background or ignoring the unavailable mounts.)

CC: (none) => ueberall

Comment 2 Colin Guthrie 2012-06-02 13:37:41 CEST

If you wish the mountpoint not to stall the boot, simply include "nofail" in the boot options. systemd is a lot more strict at adhering to these options than previous systems. If an entry does not have a nofail option, system will drop to the emergency shell if it fails to mount (as you would expect really from the description of nofail option in man fstab).

Likewise, if you use a network filesystem, you should mark it as such with the _netdev mount option. This tells systemd that it is not a local disks and it changes the point at which it is mounted (i.e. part of remote-fs.target rather than local-fs.target).

By default we consider any network filesystems specified as being critical for login - i.e. gdm/kdm etc. will be delayed until the network is ready and the filesystem has been mounted (or at least tried to be mounted). If you do not want the mountpoint to delay logins, simply include both of the options mentioned above: "_netdev,nofail".

Hope that explains things. I've updated the errata mentioned above:
https://wiki.mageia.org/en/Mageia_2_Errata#Boot_fails_when_webdav.2C_sshfs_etc._entries_exist_in_fstab

Comment 3 Colin Guthrie 2012-06-02 13:37:59 CEST

Hence closing.

Status: NEW => RESOLVED
Resolution: (none) => WORKSFORME

Comment 4 Florian Hubold 2012-06-03 23:03:48 CEST

(In reply to comment #2)
> If you wish the mountpoint not to stall the boot, simply include "nofail" in
> the boot options. systemd is a lot more strict at adhering to these options
> than previous systems. If an entry does not have a nofail option, system will
> drop to the emergency shell if it fails to mount (as you would expect really
> from the description of nofail option in man fstab).
> 
> Likewise, if you use a network filesystem, you should mark it as such with the
> _netdev mount option. This tells systemd that it is not a local disks and it
> changes the point at which it is mounted (i.e. part of remote-fs.target rather
> than local-fs.target).
> 
> By default we consider any network filesystems specified as being critical for
> login - i.e. gdm/kdm etc. will be delayed until the network is ready and the
> filesystem has been mounted (or at least tried to be mounted). If you do not
> want the mountpoint to delay logins, simply include both of the options
> mentioned above: "_netdev,nofail".

Shouldn't those things be documented/communicated in a better way with more visibility so users know that, and it will not be forgotten?

CC: (none) => doktor5000