[Discuss] I, uh, deleted the wrong kernel....

[david at thekramers.net (David kramer)]

On my desktop machine (Kubuntu 16.04LTS I think), I set up /boot too
small, because at one point my distro automatically removed old kernels,
so there wasn't need for much space.  But that doesn't happen anymore,
so sometimes it fills up.  Well it turned out that when I went to delete
some of the older kernels, it turns out I was actually running one of
them, and now Grub is very unhappy.  This is dual boot with Windows,
which still works. 

I'm sending this from my laptop.  It's now like 2AM so I need to get to
bed.  I have tried a bunch of things, but clearly I didn't exhaust
Google yet.  I'm posting this in case someone who has experience in this
can just head me towards The One True Way to fix this.  I will continue
trying tomorrow though.

Grub is looking for vmlinux-4.4.0-28-generic, and I have
vmlinux-4.4.0-34-generic and vmlinux-4.4.0-36-generic still.

- I tried booting from USB, mounting root and boot, and simply
symlinking the 36 kernel to 28. I ended up at some initrd prompt.

- I tried both mechanisms at
http://askubuntu.com/questions/28099/how-to-restore-a-system-after-accidentally-removing-all-kernels
(tl;dr  mount some of the root partition then  chroot to |||it and run "|
|apt-get install linux-image-generic").  It reported endless
dependencies that couldn't be met.  I have a feeling I may need to try
this again after finding the magic command to remove anything from the
"scheduled to be installed" list.
|

|- I tried the above chroot thing, then
https://statusq.org/archives/2012/10/24/4584/ to |grub-set-default to a
kernel I *do* have, but it still looked for 4.4.0.28.  Maybe I did the
chroot wrong? Maybe I forgot a step?


I suppose the nuclear option would be to reinstall over the current
install without reformatting and maybe kinda it may work.  I welcome
other ideas to try.  Like I said I will continue to work on this
though.  Far from waving the white flag but close to my bed.


Thanks.