Day job: oneScrote and fucking ansible
Night job: some wiki for the architecture, but just config management for the config detail : we run the docs and it’s done. Thankfully not fucking ansible; because in any environment where there are options, it’s not fucking ansible.
“Systemd is built badly by weaponized dunning-kruger” – pros
exploit [happens]
World: surprised pikachu
33 years with Linux (kernel 1.2.13, slackware). Worked at a distro. Worked in OS security – Unix and enterprise Linux. I helped build United Linux out of the dismembered corpse suse kicked over the fence as ‘collaboration’.
Because of the validation issue in the .deb package format and others, I’m on a mixture of Rocky and Nobara.
I’m subscribed to cloudLinux’s tuxcare enterprise updates for some older stuff, and I can’t recommend it enough. It’s excellent; and if almalinux releases their sLTS distro release and actually covers it for 25 years, that will be such a coup.
I’m worried at the direction Linux has been taken by IBM and I hope it can be unfucked one day. I miss the reliable, fast boots and uncomplicated tooling before this systemd shitshow.
Ansible using SSH
The moment you discover anything else, you’re gonna be so pleased. It’ll seem so modern! So fast!
Check out gatus.
It doesn’t do SNMP. That’s … bold.
Iso27002 says I can’t run this.
West refuses
The hell it did. My country refused nothing.
Mexico? Panama? Belize? Guatemala? Brasil? Peru?
Just keep naming countries. Aside from two among like 30, the “West” condemned slavery; because of course.
I can’t watch. It’s just so much “the pot isn’t boiling; it’s just a little warmer” bullshit - I’m tired, boss - that I need to know whether this guy’s gonna fail the intelligence test before I commit the time.
Please also consider mikrotik brand gear. I’ve been told they’re especially easy to manage as one moves from manual control to something declarative like terraform(opentofu).
weird grab bag
You don’t mention whether that is adherence to the FHS or denial of it.
One of those options is, to be clear, violating a standard.
The CI turned me off though. It’s like they saw how bad GitHub is for CI and said “no notes. Just like that”
And I use the CI config a huge amount.
So lennart personally blocked the revert? Fucking on-brand for all he’s wrecked in Linux.
Is he still working at Microsoft, or was he just too special for them too?
Wow. Every word of that is frightening.
I can’t think of any case where a routine update caused trouble for me.
I second this.
Since about 2002, I’ve had a growing number of machines (ent Linux, similar to fedora) who simply patch nightly via cron. I get mail on the changes. In 24 years of yum-cron I’ve seen one error (busted deps on cobbler) and caused one error (too-old smb.conf during an update) and that’s it.
So, my fuck-up aside, the only issue I’ve seen with a yum-cron patching setup (in dev/test and homelab) in ~25 years is exactly one issue on boxes running a mainstream but not prevalent piece of software due to a packaging glitch. That’s it.
natural gas
We can start just saying ‘methane’ now. We don’t need to ‘green-wash’ it like it’s not killing our air.
it is crazy that Poettering is even considering merging this
You’ve, uh, seen systemd, right? Cmon; this is just one more section for the cancer to eat.
Consider PCLinuxOS: they’re an RPM-based mandriva (mandrake/conectiva) derivative with really great and wide compatibility in stacks without the ‘modules’ shitfest RH started after no one remembered what ‘alternatives’ was for.
They don’t use systemd, but their installation is a bit shite as it’s a “live CD” installer – they pruned out the proper templatey install that mandriva has. But so far that’s the biggest issue. If they can get off networkManager we’ll be even better off, though.
there are complete ansible-repos
Ansible is toxic. Anything else?
There are many benefits to a siloed operation that all relate to preventing overwork and scope creep.
This example is part of the cost.
We use copy and paste a lot.
Heh.