tags 509070 wontfix thanks On 18/12/2008 Christoph Anton Mitterer wrote: > I'd suggest that timeout support is readded because it could be > useful, e.g. when some script tries to map a dm-crypt device, but the > usb-stick containing the key is not inserted, but the device is not so > important for that script. > Perhaps the same applies for the boot process.
Hey Christoph, as described in several other bugreports and mailinglist discussions, it's currently impossible to implement timeouts in a clean way. Both checkfs.sh and mountall.sh from initscripts will fail for a nonexistant device that is configured in /etc/fstab. The only way to prevent broken boot process for missing filesystems is to use the noauto option in /etc/fstab. But if the filesystem isn't mounted automatically anyway, then you can use the noauto option for the dm-crypt device in /etc/crypttab as well. It's not that useful after all to start the dm-crypt mapping automatically if the device needs to be mounted manually anyway. Additionally, the problems with keyfiles and keyscripts come into mind. That's why I tag that bugreport as wontfix. And I don't see a reason to implement timeout support as long as the checkfs.sh and mountall.sh initscripts don't support some flag like 'ignore-missing' to /etc/fstab. greetings, jonas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
