TIFU by rebooting before rebuilding my initfs
TIFU by rebooting before rebuilding my initfs
I had a super fast but small SSD and didn't know what to do with it, so I was playing with caching slow spinning LVM drives. It worked pretty good, but I got interrupted and came back a few weeks later to upgrade the OS. I forgot about the caching LVM, updated the packages in preparation for the OS upgrade, then rebooted. The LVM cache modules weren't in the initfs image and it didn't boot.
I should know better. I used to roll my own kernels since Slackware 1.0. I've had build initfs images for performance tweaks. Ugh!
Where's my rescue disk?
Rescue disk 🤣 It's hard enough to find a drive much less a disk.
Next time keep your old kernel around a while, you can always boot it to fix a goof instead of messing around with rescue images.
🤦♀️ I've never considered this, but it's the simplest solution and makes perfect sense. I'm always so diligent to keep my system clean to save a few megs.
This particular server is an old PowerEdge server I'm using to learn server stuff on and a practice home lab. Unfortunately, it won't boot from SD card, so I have a few DVD RW's in a drawer. I've read that there's a SD slot inside that you can emulate a floppy, but haven't explored it.
A Gentoo LiveUSB and some chroot would save your day!
I have the Debian netinst disk, but it doesn't include the dm-cache modules, so I downloaded the live DVD last night. I only get about an hour a day to work on stuff.
Aren't you supposed to add modules by putting them in some config file so they get added automatically?
Fixing your problem should also be achievable from single-user/rescue mode too, no need for a rescue disk.