Kexec: Boot a new Kernel without Power Cycling

I’m not sure if Mojeek uses Linux anywhere.

I came across Kexec today which allows you to boot a new kernel without power cycling the hardware. It is described as a developer tool. But ByteDance uses it to minimize downtime.

Mike

1 Like

As a little bit of info to yourself here, we use debian/linux everywhere, and I will soon be switching over full time when my currently-on-order 12th Gen Intel Core Processor Framework arrives. It has been a good month since they sent me a warning they’d be taking the money… :person_shrugging:

1 Like

Debian’s stability has advantages for infrastructure. I’ve been having an ok time with Manjaro at home. And it has been helpful to use Snap and AppImage containerized apps to work around the delay built into Manjaro’s testing pipeline.

I hope the laptop arrives soon. :pray:

1 Like

Thanks a lot, really I’m just sad to be looking at 2TB of NVMe and 32GB of RAM that has arrived way ahead of schedule.

The machine I have next to me that is now my server (Pinebook w/ Armbian + Open Media Vault) used to run Manjaro and I enjoyed it. I am still very much unsure about where to go when it comes to flavour.

Debian powers the rest of my house, the media stuff, the server, the flashing LED lights!

1 Like

I didn’t know anything until I bought a second hard drive and tried the top distributions listed on DistroWatch. That includes busting my Gentoo installation during an update and buying a copy of OpenBSD which included posters and music.

It was a lot easier when hard drives were swappable.

I suppose the way now is through virtual machines.

StarTech SATA Drive Bay and Caddy

1 Like

Ridiculously useful, thanks. I’m sure there will be a lot of VM-ing as I have things like Kali and Qubes that I like to use, but prefer to do so this way. In other news:

Someone from Framework clearly is watching us :eyes:

1 Like
1 Like
1 Like

Avoid kexec if you don’t need it: it opens new vulnerabilities, and is better left disabled for most use-cases. Redundancy and failover should eliminate the need when hosting a typical service. Even without the risks, there are other reasons to reboot: updates to shared libraries, SELinux policies, and init systems often warrant one.

One use-case that benefits from kexec is pubnix systems with many logged-in users. If you need to apply an unscheduled security patch while causing minimal disruption, then live-patching makes sense.

With some boot optimization, I can hit 99.9% uptime even if I reboot 2-3 times per week on Fedora. With failover, these reboots should have negligible impact.

POSSE note from https://seirdy.one/notes/2022/10/23/kexec-considered-overkill/

1 Like