Re: Installing Debian On Chromebook Pixel
Quoth Bob Proulx:
> Wow. 2560 x 1700 at 239 PPI touchscreen is quite impressive! I can
> see the attraction. Even though it is widescreen it has some vertical
> pixels.
Yeah, I dislike widescreen as an aspect ratio a lot, and the 3:2
ratio suits me well. And the high DPI makes for some very pleasingly
crisp text. Mmmm.
> They cheated and didn't provide a battery to back up the CMOS!!?!
>
> > I read "battery-backed CMOS" as meaning the CMOS settings were
> > backed by the main battery. But perhaps I'm wrong.
>
> No. You have it right.
Thanks for rereading the links for me and confirming that.
> Not present. Wow. I can't believe they designed it without any
> completely non-volatile storage of any kind.
It is pretty bad, isn't it? I suppose it's just something they
overlooked, given that it would never be noticed with their primary
usecase (user boots Google signed operating system).
> Would make me want to open the box up and find some
> place to tuck a CR2032 into it somewhere and wire the power up to
> whatever it uses for CMOS ram.
That'd be a nice idea. Though opening the box up and digging around
for the CMOS ram bit would be non-trivial, not to mention (I
imagine) actually finding space in the case.
I'm taking it as an extra prompt to ensure my backups are regular
and complete. I like to think of the company I read about quite
recently who have bots designed specifically to cause certain parts
of their infrastructure to fail at random, to ensure that their
failover systems all work reliably (I forget the company or where to
find the reference, unfortunately). I'm just going to pretend that
Google created this bug deliberately to encourage me to get my
backup procedures completely solid ;)
> The machine has an SD card slot. The main storage is only 32G. I
> think I would buy a 32G SD card and install it in the slot full time.
> Use the internal storage just as a shim to boot over to a 32G SD card.
> Then you would be safe.
At present the SD card mounting doesn't survive suspend properly
(needs to be umounted then mounted), so I'm not sure that this would
work. But it could be my specific kernel (well, the 3.10 series), or
something to do with me just calling pm-suspend straight, rather
than from a desktop environment, I don't know. Any thoughts as to
that? Should I document that (and the workaround I added to
/etc/pm/sleep.d) in the wiki page, do you think?
As I say, my solution is just backup lots and trust in my script
that shuts the computer down if the battery is below 10%. Not ideal,
but good enough.
Thanks for your thoughts.
Nick
Reply to: