I’ve had my enterprise-distro linux machines updating by cron for 22 years. I had two glitches in those 20 years, too, just like you. But in addition to my two glitches - I had to bring in one unlisted dep for cobbler and also correct the smb.conf’s old format on another box - in 20 years, I also got
- out-of-the-box
- do-nothing patch runs
- trivial back-out if I needed it
And while I know your numbers are excellent, I simply haven’t had to DO ANYTHING since deploying some boxes. They patch, they bounce later on a weekend if they need it (‘needs-rebooting’ is centralized because ALL software installs are) and I can patch while under load because linux write-locks instead of read-locking. My effort is to check ‘some time later’ and ensure things are working in ways nagios doesn’t catch.
Printer issues? Nah. Supply thing. App not working because java/perl/python/DLLs rug-pulled a dependency? Proper packages list hard dependencies, so that cobbler thing is a bug not an expectation. Network offline? nah. Reboots? timed at 3 minute downtime (1 min before systemd), or 7 minutes if I just updated 1gb of gitlab install because it starts like a manatee.
It’s really a different world; and while I’ve teased the heck out of my windows peers, it’s a true statement.
HeyJoe@lemmy.world 2 weeks ago
Absolutely, I don’t disagree with your statement at all. I work heavily in systems administration and recently transitioned to networking. I deal with Linux systems, servers, vm’s, Azure daily, and for stability nothing beats Linux. I just tend to agree with the statement above commenting on how you always see these Windows articles, yet almost none actually affect you in the end.