On Lu, 11 aug 14, 23:53:54, Michael Biebl wrote: > Am 11.08.2014 23:35, schrieb Andrei POPESCU: > > I have been thinking about this incident since then and watching all the > > "system doesn't boot due to fstab errors" reports (I'm subscribed to > > both -user and -pkg-systemd-maintainers). > > > > I'm not sure it's a good idea to just fail the boot completely whenever > > there are errors like mine in fstab, but instead it would be much > > friendlier if systemd would try to bring up as much of the system as > > possible and start the emergency console only if e.g. / and /usr are > > missing. > > > I already commented on that. I think it's wrong to just continue booting > and starting potentially critical services. > How is systemd supposed to know which file systems are vital? It can't, of course (nor should it), unless receiving enough information from the admin, but see below.
> You might have database on separate partitions and you might risk losing > data if the services are started but the partitions aren't available. I understand these concerns, however: 1. it seems very brittle to me that a wrong fstab entry is capable of blocking the entire boot. This will be mitigated once emergency mode is fixed, but still won't be much comfort for headless systems. 2. sysv-rc would not stop either, so this would not be a regression Another possibility would be if admins could designate specific filesystems as critical for booting with a fstab parameter like 'bootwait', where systemd stops if that particular filesystem couldn't be mounted. Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser Offtopic discussions among Debian users and developers: http://lists.alioth.debian.org/mailman/listinfo/d-community-offtopic http://nuvreauspam.ro/gpg-transition.txt
signature.asc
Description: Digital signature
