Hibernate and lack of docs

[Fred Smith]

On Mon, Nov 10, 2014 at 09:42:07AM -0700, Pete Travis wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 11/10/2014 06:36 AM, Martín Marqués wrote:
> > 2014-11-09 20:31 GMT-03:00 Pete Travis <lists at petetravis.com>:
<snip>
> > I had a Dell laptop a which I bought about 4 years ago. I almost
> > always suspended it (no hibernation) and when it got to 14 months
> > (warrant was over) it died (hard die, like in motherboard chip got
> > fried).
> >
> > The guy in Dell told me not to suspend the laptop, at least not for
> > long periods of time, which really surprised me.
> 
> Post hoc, ergo propter hoc?  I've had Dell laptops from then, older,
> newer.  Suspend triggered by lid close was the normal disuse state for
> all, with periodic hard halts due to battery depletion :)  Generally,
> some laptops fail, and some don't - unless there's an explicitly clear
> impetus, ie prolonged dirty power or physical damage, there's a lot of
> speculation involved.  It seems to be more common with machines equipped
> with heat-producing dedicated graphics, so I avoid them.
> >
> >
> > About cold boot, well even if the cold boot is fast, there is lots of
> > things I need to get starting before I start to work (ssh keys, login
> > to monitoring systems, etc) which make a cold boot extremely tiresome.
> >
> >> On point 3, one would hope that resuming from hibernate didn't require
> >> you to manually edit the grub configs every time :)  I'm not well versed
> >> in the area, but it smells like at least one bug/deficiency  from here.
> >
> > This looks like a kernel bug, as you can specify a boot partition
> > using UUID, so why not the one to resume from.
> >
> >> [1]
> >>
> http://docs.fedoraproject.org/en-US/Fedora/20/html/Power_Management_Guide/pr01s02.html
> >
> > Thanks.
> >
> Sure, just because hibernation has 'fallen out of favor' doesn't mean
> you shouldn't expect it to work :)  The kernel devs are good about
> triage, filing a bug there would be a good start even if the kernel
> isn't directly at fault (systemd, maybe?)

Agreed, it SHOULD work, wherver possible (within constraints of weird/
broken hardware).

I sometimes need to shut down a machine (laptop or desktop) temporarily
while I have important stuff open or otherwise in progress, and do not
want to lose it. Suspend isn't always suitable because you can't unplug
a desktop to move it while in suspend, unless you don't mind losing
what you're doing. And of course, if I NEEDED hibernate, suspend just
wouldn't do.

As a programmer myself, I understand (painfully, too often) that for
any arbitrary situation one might find oneself stuck in, that some of the
more attractive choices just won't do, and the really useful choice can
be painful, so I DO have sympathy for people who would like to deprecate
hibernate because the hardware situation is a big mess.

-- 
-------------------------------------------------------------------------------
 .----    Fred Smith   /              
( /__  ,__.   __   __ /  __   : /     
 /    /  /   /__) /  /  /__) .+'           Home: fredex at fcshome.stoneham.ma.us 
/    /  (__ (___ (__(_ (___ / :__                                 781-438-5471 
-------------------------------- Jude 1:24,25 ---------------------------------