My primary workhorse right now is an HP Omen running Fedora 22. Thinking and talking about how to restore my work environment from a total failure, I came to the conclusion that there are three primary things that need to be preserved:
Item 1: My entire user directory under
/home, since that’s where all my stuff is.
Solution: Crashplan is impressing me – slick commercial-quality UI, completely free for mounted “local” disk backups and peer-to-peer backups through their central broker. Naturally they also offer managed off-site backups for a cost, but they don’t ram it down your throat. I think they’ll be getting some of my business.
Item 2: A list (preferably in Kickstart format) of user-installed software packages, so that I can restore all of the apps I use.
Solution: this is very distro-specific, but
yum’s heir apparent) has a kickstart-friendly report that I will integrate into a once-a-day kickstart generator script:
dnf history userinstalled
Item 3: A snapshot of everything under
/etc, which tells the computer how my apps and service should be run.
Solution: I found an awesome utility called etckeeper that turns
/etc into a git (or mercurial, or svn) repo and can automatically push changes out to multiple targets. On its own it is nice, but the dnf-etckeeper packages hooks this behavior directly into the package manager – so, every time I run dnf for an update or a new package, my
/etc snapshot is updated as well.
Thinking about this now, I realize that for completeness I need to copy
/var/spool/cron/* somewhere as well. The only cron job I’m running is the tail-end of an IFTTT hack to gets photos off of my iPhone and into shotwell, but that’s fodder for another post.