Bug#780364: [Pkg-sysvinit-devel] Bug#780364: The problem disappeared, when I removed /etc/adjtime file and rebooted

[Dmitry Bogatov]

control: tags -1 +moreinfo

[ Added initramfs-tools maintainer to thread ]

[2015-05-05 14:19] Michael Biebl <email at michaelbiebl.de>
> [...]
>
> Most likely, /etc/adjtime was configured to use local time and you ran
> into [1]. In systemd, we avoid that error by skipping the file system
> check if already done in the initramfs [2].
> I think sysvinit should do the same and test for the flag files created
> by initramfs-tools [3].
> [...]

Thank you. Seems reasonable.

Dear co-maintainers, I seek you advice how best approach this situation.
According to initramfs-tools(7), if /run/initramfs/fsck-{root,usr}
exits, corresponding partition was already checked and should be
skipped.

 * despite name, `checkroot.sh' also does another function: it enables
   swap. So it sounds reasonable to factor (there is quite a logic) swap
   manipulation code into another init script (swapon.sh, for example).

   But since currently `status' of `checkroot.sh' return 0 if either
   swap on or / is rw, such refactoring would slightly change semantics
   of `status' command. Would it cause breakages?

 * checking of /usr in `checkfs.sh' is performed by `fsck -A'. Any
   ideas, how should we tell it to skip /usr?

   One possible approach is that if sixth field in `/etc/fstab' is `0',
   `fsck -A' will ignore corresponding file system, but who have
   responsiblity to set it?