Remember my F20 nvram issues with Lenovo x120e? More sagas

[Adam Williamson]

On Fri, 2014-05-30 at 14:40 -0400, Robert Moskowitz wrote:
> A quick piece of  review then on to day's failures.
> 
> Back in December?  I was working on installing F20-64 on my Lenovo x120e 
> and could not update NVRAM.  You here helped me get a working system.
> 
> ============ now on to recent history =============
> 
> Then the audio started getting flacky and could tell if it was hardware 
> or a software update (headphones worked, but not built in speakers).  So 
> I picked up another x120e from ebay that had an oem Windows7 on it.  I 
> pulled that drive and put in my drive from the old system.  I figured 
> that since it was working without nvram info it would work in another 
> box.  And it did.  For a while.
> 
> Then one reboot after a kernel update, it would not reboot.  A long set 
> of messages and then failure to sync drive and powering off.  I found 
> that if I hit s or <ctl-s> a lot of times at the right point it would 
> boot up.  I rebooted as little as possible and only applied updates when 
> gnome would freeze up (in a tty session). Finally it got too bad, and I 
> put the drive back into the old system with the flacky audio where I am 
> right now.
> 
> ===================== Finally today's failures =================
> 
> So I ordered a new SSD drive and today set about installing F20-64. See 
> bug 975537, and no, Adam, I did not save anything.  Instead I set about 
> doing an i386 install!  That seemed to go ok until I got to the reboot.  
> It goes through a reboot and shuts off.  I have tried with <alt-d> to 
> catch any messages, but none that I can see.  :(
> 
> So now what?  Adam, I CAN do another x64 install and if you give me the 
> steps to rescue any logs (you mean they were not part of the upload?) I 
> will attach them to the bug report.

That would probably be best. The logs get attached automatically when
libreport files a *new* report, but not when it thinks your bug is a
dupe - unfortunately, any time bootloader installation fails in a UEFI
install it tends to winds up as a dupe of some older existing bug,
because of how the error gets reported.

So, what you can do is boot a live image, mount the installed system
that you can't get to, and pull /var/log/anaconda/program.log (or
anaconda.program.log, I forget what it winds up being called) out and
attach it to the bug report, or to a *new* bug report.

> Or I COULD do a rawhide install.  Point me where to get the ISO image 
> and I am willing it give it a go.  I DO at least have this system to 
> work from.

I updated https://fedoraproject.org/wiki/Rawhide recently to try and
make this sort of thing clearer, can you check that and see if it works
as a useful guide?

> thank you for your support.  Now I think I will try that Fedora 20 arm 
> install again.  This time recable my monitor so I have direct hdmi 
> connection to get me through firstboot.
> 
> 

-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net