On Thu, 2014-10-02 at 23:04 +0100, Ben Hutchings wrote: > On Thu, 2014-10-02 at 22:26 +0100, Ben Hutchings wrote: > > On Thu, 2014-10-02 at 18:39 +0100, Ben Hutchings wrote: > > > On Mon, 2014-09-29 at 01:44 +0100, Ben Hutchings wrote: > > > > Currently initscripts will not work correctly if /usr is mounted by > > > > initramfs-tools. Firstly, fsck will fail and the boot process will > > > > stop. If you press Ctrl-D to continue, the boot process will succeed > > > > but /usr will not be remounted read-write if necessary. > > > [...] > > > > > > Also, we must not run fsck in the initramfs because the system clock may > > > not be correct yet; see #763777. > > > > > > Unfortunately, systemd does not fsck /usr if it's already mounted. > > > Which means this is now broken for all init systems! > > > > It looks like we can sort out the system time just by adding hwclock and > > some config files into the initramfs. hwclock already hooks into udev, > > and the kernel ensures that the adjustment of system time to a local > > time zone is done only once. So please assume that the initramfs will > > continue to fsck /usr. > > Currently this works fine with systemd, but util-linux will need a small > patch to avoid double-adjusting the clock when booting with > sysvinit/initscripts.
I' made an NMU to util-linux to avoid double-adjustment (2.20.1-5.11). The necessary change is also on the 2.25 branch. I've just now uploaded initramfs-tools 0.118 which includes these changes, many more fixes, and a versioned dependency on util-linux. Ben. -- Ben Hutchings One of the nice things about standards is that there are so many of them.
signature.asc
Description: This is a digitally signed message part
