System Update froze - UbuntuMATE 16-04

[Colin Watson]

On Thu, Jun 21, 2018 at 09:43:16AM +0100, Colin Law wrote:
> On 21 June 2018 at 09:32, Bret Busby <bret.busby at gmail.com> wrote:
> > On 21/06/2018, Colin Law <clanlaw at gmail.com> wrote:
> >> sync
> >
> > After about fifteen minutes, no displayed response - cursor goes onto
> > next line after the line in which the command is entered, an the
> > cursor sits there flashing at me, in an otherwise blank line.
> >
> > So, the command "sync" also appears to have been hanged (and, yes, I
> > do know the difference between hanged and hung :) ).
> 
> So it is that bug that you are hitting.  The update runs sync to flush
> data to disc and that cannot complete for some reason.

Well.  Yes, Bret is hitting the initramfs-tools bug you linked to above.
However, that only happens when the system is already in a bad way (in
that sync never completes), and *that* is not caused by initramfs-tools.

Investigating what's causing the sync hang would involve poking through
system logs.  /var/log/syslog is the first place I'd look if I saw this
on one of my systems, as there are multiple possible causes: as well as
things like stuck remote filesystem mounts or whatever, it could also be
due to incipient disk failures that would cause the kernel to produce
alarming log messages.

> > It surprises me that this bug original report is apparently 16 months
> > old, and, apparently a critical bug, and, since then, two "point
> > releases" and a new version (18.04) have been released, with this
> > critical bug not being fixed, with the bug appearing to take 16.04
> > from being a "stasle" release, to a status equivalent to Debian sid,
> > with all that that name involves.
> >
> > I am now concerned that the bug may also apply to 18.04, as the bug
> > has apparently not been fixed.
> 
> Yes I believe it does.  I guess the reason it has not been fixed is
> that it is extraordinarily rare, as can be seen from the small number
> of reports compared to, I guess, many many millions of times kernel
> updates have been run.

I would describe the initramfs-tools bug as "indirectly critical".  It
causes what might otherwise be an ignorable problem (in the case of a
stuck remote filesystem mount) to be promoted to an upgrade failure, and
that's certainly no good.  But, as you say, the underlying cause here
could well ultimately be a hardware failure, and not something that
Ubuntu can in fact fix beyond just reacting to it slightly more
gracefully.  We don't know yet.

-- 
Colin Watson                                       [cjwatson at ubuntu.com]