On 2013-04-15 14:01 , Daniel Baumann wrote: > lxc cannot know if the user wanted it ro or not before the container has > been started into the system entirely, especially since there's a wild > use of aufs or overlayfs with (shared) ro rootfs between containers.
I agree that the *rootfs* might have that situation. But note that the config dir apparently has to be read/write too. The error/warning messages/notices that are shown when one starts with 'lxc -n <name>' are all related to the network interface. There is no warning to the effect that a read only filesystem is causing this failure as apparently some state cannot be properly written and one thus is looking at network errors but that is not the cause of this. Finding out that the ro flag is the problem is thus quite difficult and one just has to look in the right location for it to be found. As such, it would be really nice if there is an error/notice shown that the filesystem is read-only and that that causes the error. Hopefully this bug will be found by people who google for related issues though... Greets, Jeroen -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
