@rozzin The devices perform some kind of auto-negotiation to decide which way the power/data flows. But like everything automatic, it often fails. And devices rarely have a manual override.The most popular incident was probably with the first Type-C Macbooks in 2015. If you plugged an Android phone into them with a Type-C cable, the phone would chargâĻ
I'll admit Linux isn't quite as easy to use as Windows, you still have to paste stuff into a terminal once upon a blue moon to troubleshoot something. But in general, everything can be done with some graphical application these days.
Linux is definitely easy enough for any regular person to use now. At least almost as easy as Windows is. Users are simply unwilling to even attempt the learning curve.
It's logical, it's the same reason that most regular people are unwilling to switch between iOS and Android. If they like their Samsung phone, they'll keep thinking they're easier to use than iPhones. But if they've used iPhones, they'll keep thinking that one's easier.
Users simply hate change. Which is sad, but fairly obvious if you've ever dealt in IT support. Turns out that most people learn by memorization, not by logic.
First it's the stuff about people being pissed about a woman taking the front in a war game.
But now in the beta, people are discovering that the profanity filter is absolutely broken. "White Man" is apparently profane, while "Black Man" is not.
You can't say "DLC", but if you specifically say "Free DLC" then we're all good.
Apparently you also can't say the word "Nazi". In a game ABOUT FREAKING WORLD WAR 2
12:08 - I was back on a computer so I started the server back up. Seemed to reboot fine
15:54 - Server went down again, for a still unknown reason
~16:15 - Got home and started troubleshooting the server while preparing/eating dinner. Gave no decent results. Box didn't even POST via the KVM
17:04 - Gave up troubleshooting, opened ticket with host
17:12 - Ticket assigned to technician
17:12 - Service attempt by DC technician #1 17:16 - Service attempt by DC technician #2 17:25 - Service attempt by DC technician #3 17:27 - Service attempt by DC technician #4
17:31 - Ticket: "Hello, we're sorry to inform you that blah blah blah", AKA "Your server is fucked buddy. Want a replacement?"
17:35 - I approved server replacement
????? - Oh right, I have one of those limited edition boxes so finding a replacement isn't easy
19:24 - Tech confirmed they'd get it replaced one way or another
20:30 or something - New server was available
21:30 - I did a borg mount and started copying over everything I needed so I could start working at it during downtime the next day
** Next Day **
06:30 - Started copying the database dump as well
09:00 - Started setting up Mastodon again. Turns out Mastodon is gay, who knew?
10:30 - Literally only installed/compiled gems and stuff for the past hour, fml
It just flat out doesn't. Background tasks alone will saturate the whole link and just trying to ping something times out.
2-3 applications with push notifications or more and it's actually flat out impossible to access the internet over 56k using an Android phone thanks to it saturating its own link.
@lain Also have you tuned GitLab. I recommend you lower unicorn workers to 2. Set sidekiq workers to only a couple (Instead of the default overkill 25 or whatever it was). And disable the Prometheus monitoring plugin.
You often shave of a nice gig of ram simply by doing that.
When someone changes a damn salad emoji to "include vegans" and you can't even tell if it's a joke or if they're dead serious, we've definitely gone down the wrong path.