But why?

I a world where we can’t really be sure what’s in an upgrade, a super-clean start that burns any ephemeral data is about the best way to ensure a consistent start.

And consistency gives reliability, as much as we can get without validation (validation is “compare to what’s correct”, but consistency is “try to repeat whatever it was”).

breakeven

Not a word, my dude.

able to open their browser and look something up without immediately getting distracted by the 6 tabs

That’s the methylphenidate talking.

Oh noo; forgejo is still not connected? I had high hopes. What’s the timeline?

(Honestly, but for the wonky CI spec language I may have switched already)

Company A permits federation of project 2 with contractor B who agrees. Oh look! No need to add 21 people to your AD. Contract done? De-fed.

Google feds with GH for AOSP dev, because it’s 2023. Users don’t need to even know where the repo is hosted or whether the real meat is another hop inside.

Company “BCFerries”, an imaginary organization, happens to run the largest fleet of mobile DCs in the country, with each mobile DC being 6 HA racks, three on a side, dehumidifiers o-plenty. Engineers stationed aboard need to lob tickets and hot fixes on the go, and sub them for review when the mobile DCs get a good link, 30 min out of every 2 hours. Roaming 2/2/2tb node swaps spit with the stationary nodes when it smells the VPN, and then gets ready to go again.

Repeat that above, but say ‘Maersk’. I’m betting evergreen/evergiven is on VSS.

Enough examples?

Running a federated GL is conceivably a set-it-and-forget-it like a lot of federated stuff already is, and you debug the glitches and patch like normal.

Given the 5GLs I still run were all installed by VMware/terraform/chef/RPM, patched automatically with package promotion and watched for anomalies, it’s already negligible effort. Double nothing is …let’s see …carry the 0 …integrate the square …nothing.

HELL NO.

“May federate” doesn’t necessarily mean “must federate.” Your concerns could be met if they include the standard kill switch in gitlab.rb .

(Now show me the kill switch for the bloated crappy web editor)

I saw this and was pleased, actually. The relative opacity of containers makes them a validation challenge and hides versioning from standard tooling used for large host populations and/or enterprise.

Even if they sparkle.

Docker-dependent? It looks fantastic, but I have no containers in my home-lab – and it’s based on my time managing OS security for an OS. I’m stuck living vicariously through the rest of you, so report back often.

Postiz is a social media scheduling tool supporting 19 social media channels:

Thank you for including this summary. It’s surprising how often that’s left out and it’s always valuable.

Sadly, it’s container-dependent, it seems. Also, it’s asking for supply-chain exploits and violating ISO27002 with pnpm, but for a PoC setup it looks excellent.

If there’s an article with words, can you pass it along? I’m not inclined to go on a Google website and listen to someone slowwwwly tell me why Google is bad.

Life’s too short to listen to people who talk slow.

Yes, technically it’s still on the tiny screen of your phone with a really bad ux to make it go, sure. But you see how ‘went away’ is both true for web and effectively true for phone app? You do see that, right?

Pixelfed? That’s an IG emulant. What about Friendica? Do they have groups?

They have a really good racket going on and want to make sure people keep giving them way more money than is necessary. It’s simply not true that Infosec.Exchange requires $5000/month to operate unless they’re doing something very wrong or just straight up lying.

Yes, internet rando. I totally believe your solid calculations based on … vibe?

I support ads.

Oh, calm down. I don’t support the ad level of Facebook, nor the targeted ads, nor the algorithm.

And we, as users, get to decide when too many ads are too many, with our feet.

No. Their reward for having users is that they’re in control. Expecting users to then pay them for that control is fucking stupid,

You DO realize that not everyone works to attain power over other people, right?

but I don’t expect most people to realize it.

The reason people don’t realize that site owners’ reward for forking over half a salary in hosting costs for some nebulous power to hold other people in their clutching fists and cackle maniacally is because that’s not the motivator here.

I look forward to when you can see that.

Whoo-wait…nope-oooosh

…not for my OS.

Slop, posted by Ai? Is there a slopper even involved, or did another Ai algorithmically decide what the most click-baity prompt should be?

Please, for the love of god, look at other things instead of Ansible.

Definitely do openTofu for infrastructure and deployment, but for configuration of VMs please learn about puppet, saltstack, chef(cinc.sh) and especially mgmtConfig .

Ansible, by comparison, better matches what we were doing in 2002 at 1/10th the speed, and it’s like pascal levels of wordy.

Learning about options and finding one that works well for you will often give you a much better experience than fucking Ansible.

If you do abandon all hope, though, then go ahead and do Ansible; but remember if you do: there are better options, and hating Ansible doesn’t mean you hate automation.

Is immich still container-dependent?